Tool Directory¶
Artificial intelligence (AI) tools for teaching, study, research, and clinical work, and where the American University of Antigua (AUA) stands with each.
Find Tools by Task¶
Each task lists its tools; the names link to their cards in the directory below.
Where to Start notes are suggestions, not endorsements. Each is picked because its description here, or a guide on this site, says it fits the task; not a ranking or an AI Committee review.
Ask and Draft (12): ChatGPT, Claude, DeepSeek, Gemini, Grok, Kimi, Lumo, Microsoft 365 Copilot, Perplexity, Qwen Studio, Vibe, Z.ai
Where to Start
- Answers you can check against sources: Perplexity (freemium). It answers with citations to live web sources.
- A standing setup for a course or a role: Claude or ChatGPT (both freemium). Both keep your instructions and files in a Project, which Standing Setups walks through.
General chat assistants. Choosing Your Interface explains which kind of tool fits which work, and the prompt library holds reusable templates.
Microsoft 365 Copilot is Licensed: procured by the university.
Research and Literature (17): Asta, Claude Science, Consensus, Covidence, Elicit, Gemini Notebook, Google Scholar Labs, OpenEvidence, Prism, Rayyan, ResearchRabbit, SciSpace, Scite, Scopus with AI, Semantic Scholar, Undermind, UpToDate Expert AI
Where to Start
- Searching and mapping a topic: Scopus with AI (AUA-licensed). AI for Research starts here for searching, orientation, and citation chasing.
- Screening and extracting papers: Elicit (freemium). It pulls study details into tables you can check column by column.
AI for Research maps these tools to each stage of a project, starting with Scopus with AI, licensed through the AUA Library. The Reviewing the Literature playbook walks one review end to end.
Scopus with AI is Licensed: procured by the university.
Study for Exams (13): AMBOSS, Aquifer, Gemini Notebook, Medical Student AI, Nabu Tutor, Neural Consult, Oncourse AI, Osmosis AI, SimChat, StepGenie, StudyFetch, Thea, UAsk
Where to Start
- One notebook per course: Gemini Notebook (freemium). It answers from the slides and notes you upload, with citations.
- Practice from your own slides and notes: Neural Consult, StudyFetch, or Thea (all freemium). Each builds questions or flashcards from what you upload.
The students page routes the study workflow, and Standing Setups builds a question tutor from your own lecture notes.
Make Slides or Posters (12): Beautiful.ai, BioRender, Canva, ChatGPT Work, Claude, Claude Design, Gamma, Manus, Mentimeter, Microsoft 365 Copilot, Napkin AI, PowerPoint Design Suggestions
Where to Start
- Make a new deck: Claude (freemium). Its Claude Slides feature, on every plan including Free, drafts the deck from notes, a report, or a conversation and exports it to PowerPoint, PDF, or Google Slides.
- Improve or restyle a deck you already have: PowerPoint Design Suggestions (needs an organization's license), Canva (freemium), or Claude Design (freemium). Design Suggestions works inside the .pptx file itself; Canva and Claude Design import the deck and export the result back to PowerPoint.
To start a new deck, Claude Slides builds one from your notes or a conversation. To improve an existing PowerPoint deck, start with tools that work on the .pptx file itself. In Claude, four document skills also build decks and documents directly.
Microsoft 365 Copilot is Licensed: procured by the university.
Polish Writing (3): Grammarly, Paperpal, Writefull
Where to Start
Grammar, clarity, and language feedback. For feedback on student work, the Feedback on Student Writing playbook shows a method that keeps the student's identity out of the tool and the judgment with you.
Transcribe Audio (5): Fathom, Microsoft 365 Copilot, Microsoft Teams Transcription, Otter, Whisper
Where to Start
- Recordings you would rather keep on your computer: Whisper (free). It transcribes audio on your own computer, so the audio never leaves it.
- Teams meetings: Microsoft Teams Transcription (needs an organization's license). Live transcription is built into Teams under the institutional Microsoft 365 license.
Whisper runs on your own computer, so the audio never leaves it. Otter and Fathom transcribe and summarize video calls.
Microsoft 365 Copilot is Licensed: procured by the university.
Use My Own Documents (3): Cowork, Gemini Notebook, SciSpace
Where to Start
- Questions across a set of your files: Gemini Notebook (freemium). It answers from what you upload, with citations back to the passage.
- Research papers: SciSpace (freemium). It answers questions over uploaded papers and can extract their data into tables.
Gemini Notebook answers from what you upload, with citations, and Standing Setups makes any assistant remember your course.
Make Images (7): Adobe Firefly, BioRender, ChatGPT Images, Flow, Ideogram, Midjourney, Napkin AI
Where to Start
Generated anatomy looks right and is often wrong. AI-Generated Images in Teaching shows how to check a figure before it goes in front of a class.
Make Video (7): Adobe Firefly, Flow, Hailuo, HeyGen, Kling, Runway, Synthesia
Where to Start
- A presenter video from a script: Synthesia (freemium). It turns a script or document into a narrated presenter video without filming.
Short clips from text or images, and presenter-style avatar videos from a script.
Make Audio (5): Adobe Firefly, Adobe Podcast, ElevenLabs, Suno, Udio
Where to Start
- Narration for course material: ElevenLabs (freemium). It reads written material aloud in a generated voice.
- Cleaning up a recording: Adobe Podcast (freemium). Its Enhance Speech filter removes background noise and echo from voice recordings.
Music, narration, and cleanup of recorded speech.
Hand Off a Task (11): Antigravity, ChatGPT Work, Claude Code, Codex, Comet, Cowork, Dots, Manus, Microsoft 365 Copilot, OpenClaw, Perplexity Personal Computer
Where to Start
- A first agent task: ChatGPT Work (freemium) or Cowork (paid). AI Agents suggests either one, on a task whose result you can fully check.
AI Agents is the field guide, and Your First Agent Session runs one in 20 minutes on a folder that cannot be hurt.
Microsoft 365 Copilot is Licensed: procured by the university.
Keep Data on My Computer (5): Jan, LM Studio, Ollama, Open WebUI, Whisper
Where to Start
- Chat with no command line: LM Studio (freemium). Running Models Locally calls it the simplest start.
- Transcribing recordings: Whisper (free). The audio never leaves your computer.
These run on your own machine, so what you type or record stays there rather than on a vendor's servers. Running Models Locally is the walkthrough and Hardware for Local AI sizes the machine. Patient and student records remain covered by the privacy section of the AI Responsible Use Policy wherever they are processed.
See What AUA Licenses (2): Microsoft 365 Copilot, Scopus with AI
Licensed means institutionally licensed or procured by the university; this list grows as the university licenses more. A license on its own is not an approval for patient or student records; the AI Responsible Use Policy is the reference for those.
All 2 are Licensed: procured by the university. 2 more carry the Institutional cost label, meaning they need an organization's license: Microsoft Teams Transcription and PowerPoint Design Suggestions.
Unless a card shows another status, a tool is Listed: in the directory, not reviewed by the AI Committee. What the statuses mean
Browse All 84 Tools by Category
Assistants¶
Assistants: 12 tools · 1 Licensed
General-purpose chat assistants for questions, drafting, and analysis.
ChatGPT¶
Status: Listed
OpenAIfreemium
Widely used artificial intelligence chat assistant for drafting, summarizing, brainstorming, and general question answering. Since October 2026, on every plan including Free, its answers can include interactive charts, diagrams, forms, and small tools such as a calculator, a feature OpenAI calls Intelligent UI (user interface).
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Claude¶
Status: Listed
Anthropicfreemium
General-purpose artificial intelligence assistant for writing, analysis, coding, and answering questions through a chat interface. On every plan, including Free, its Claude Slides feature builds a presentation from notes, a report, or a conversation and exports it to PowerPoint, PDF, or Google Slides.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
DeepSeek¶
Status: Listed
DeepSeekfree
Chat assistant from the Chinese lab DeepSeek, built on reasoning models whose weights are openly published.
Operated from China, where its privacy terms say your conversations are stored, which is worth knowing before you use it for work material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Gemini¶
Status: Listed
Googlefreemium
Google's artificial intelligence assistant. Since October 2026 a free personal account gets the Gemini app with the model chosen for it, most prompts answered by the fast Flash-Lite model and some that need deeper reasoning sent to Flash or Pro; choosing a model yourself takes a paid Google AI plan. Gemini inside Gmail, Docs, and other Google apps comes with the paid Google AI plans or a Google Workspace edition that includes it.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Grok¶
Status: Listed
xAIfreemium
Chat assistant from xAI (now SpaceXAI, part of SpaceX), on the web, in its own apps, and inside X, with real-time web and X search, voice mode, image generation, and an in-chat Build mode that makes small apps and websites. Free within limits; paid SuperGrok plans raise them and add video generation.
On its consumer plans, your conversations are used to train its models unless you turn off "Improve the model" in its data settings or use Private Chat, and its privacy policy asks you not to share health information, which is worth knowing before you use it for work material.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Kimi¶
Status: Listed
Moonshot AIfree
Chat assistant from Moonshot AI, built on the Kimi K3 model family, with long-document reading and agent features.
The international version, linked here, is operated from Singapore, and its privacy terms do not say which country stores your conversations, which is worth knowing before you use it for work material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Lumo¶
Status: Listed
Protonfreemium
Chat assistant from Proton, the Swiss company behind Proton Mail, running openly published models on Proton's own servers, with optional web search, file uploads, and image generation. Free to use, even without an account, with lower limits on messages, chat history, and image generation than the paid Plus plan.
Operated from Switzerland on servers Proton controls. Proton's pages say saved chats are stored with zero-access encryption that Proton cannot read and are not used to train models, apart from feedback you choose to send; each message is still decrypted on Proton's servers while the reply is generated.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Microsoft 365 Copilot¶
Status: Licensed
MicrosoftAUA-licensed
Artificial intelligence assistant built into Word, PowerPoint, Outlook, and Teams under an institutional Microsoft 365 license. Its Cowork agent carries out multi-step tasks across email, meetings, and files and returns a finished deck, report, or email, pausing for approval before steps such as sending a message; an administrator turns it on. Word itself transcribes a recording or an uploaded audio file with each speaker separated, up to 300 minutes of uploaded audio a month on a standard Microsoft 365 license.
AUA licenses Copilot for a small number of faculty; students are unlikely to have access through AUA.
What AUA provides: access, through a university license.
Checked October 2, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Perplexity¶
Status: Listed
Perplexityfreemium
Search-centered assistant that answers questions with citations to live web sources, useful when you need traceable answers rather than open-ended chat.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Qwen Studio¶
Status: Listed
Alibabafree
Alibaba's chat assistant, built on the Qwen model family, with document, image, and voice input alongside text chat.
Operated from Singapore on Alibaba Cloud; its privacy terms say your conversations are stored there and may be shared with affiliated companies, which is worth knowing before you use it for work material. Smaller Qwen models, listed under open-weights models, run on a laptop, where nothing you type leaves the machine.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Vibe¶
Status: Listed
Mistral AIfreemium
Chat assistant from the French lab Mistral, renamed from Le Chat in June 2026. Combines a conversational mode with work and coding agent modes that connect to tools such as email, calendar, and code repositories.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Z.ai¶
Status: Listed
Z.ai (Zhipu AI)freemium
Chat assistant built on the GLM model family from Chinese lab Zhipu AI, with agent modes for slides and documents.
Operated from Singapore; its privacy terms say your data is generally processed there and may be transferred to other companies in its group, which is worth knowing before you use it for work material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Agents¶
Agents: 10 tools
Tools that plan and carry out multi-step tasks on your behalf.
The AI Agents field guide frames this category: what agents are good and bad at, what to watch for, and where to start. Your First Agent Session runs one in 20 minutes.
Antigravity¶
Status: Listed
Googlefreemium
Google's development platform built around agents: a desktop app for running and supervising agents, an editor, a command-line tool, and a software development kit, which plan and execute multi-step coding tasks using Gemini and other models. Google moved individual users of its earlier Gemini Command Line Interface to it in June 2026; that tool remains available to organizations and through paid access.
Free and paid plans (through Google AI Pro and Ultra); quotas are set by capacity and have changed several times.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
ChatGPT Work¶
Status: Listed
OpenAIfreemium
OpenAI's agent for finished work: give it an outcome and it returns documents, spreadsheets, slides, sites, or analyses. On the desktop it can use local files and applications with permission; included with every ChatGPT plan on desktop (limited on Free and Go), with web and mobile on Plus and higher plans.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Claude Code¶
Status: Listed
Anthropicpaid
Anthropic's agent for software work, available in the terminal, in code editor extensions, in a desktop app, and in the browser. Plans and executes multi-step tasks against your own files; included with paid Claude plans.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Codex¶
Status: Listed
OpenAIfreemium
OpenAI's coding agent, part of the ChatGPT desktop app since the July 2026 merge, alongside a command-line tool, a code editor extension, and a cloud service. Works with local folders, repositories, and terminals. Included with every ChatGPT plan, including Free; the cloud service, which keeps tasks running while your computer sleeps, needs Plus or higher.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Comet¶
Status: Listed
Perplexityfreemium
Web browser with a built-in agent that can summarize, navigate, and act across your open tabs and accounts. Free to use, with higher limits on paid plans.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Cowork¶
Status: Listed
Anthropicpaid
Anthropic's agent for non-coding knowledge work, on web, desktop, and mobile, and now becoming part of Claude itself (rolling out to Pro and Max first). Reads, edits, and creates files in folders you grant it, returning finished documents and analyses rather than instructions.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Dots¶
Status: Listed
OpenAIpaid
OpenAI's always-on agent inside ChatGPT: you give your dot an ongoing responsibility, and it keeps working between conversations on its own cloud computer and browser, using the apps you connect, while an automatic review decides which steps that affect your accounts or share information need your approval. Rolling out on Pro (adults outside the European Economic Area, United Kingdom, and Switzerland), Business Premium, and Enterprise (once an administrator turns it on); not on Free, Go, or Plus.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Manus¶
Status: Listed
Manus (Butterfly Effect)freemium
General-purpose autonomous agent that breaks a goal into steps and executes them in a cloud workspace, producing research, documents, slides, and simple applications.
Independent again since August 2026 after Chinese regulators blocked Meta's acquisition; existing accounts may have changed during the separation.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
OpenClaw¶
Status: Listed
OpenClaw (open source)free
Open-source, self-hosted personal agent that connects a model of your choice to the messaging apps you already use. Free software; you supply the model and the hardware.
Self-hosted, and it acts with whatever access you grant on your computer and messaging accounts. Because it reads incoming messages, text in those messages can steer it, so giving it only the access a task needs limits what a stray instruction can reach.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Perplexity Personal Computer¶
Status: Listed
Perplexityfreemium
Perplexity's desktop agent, which orchestrates multiple models across local files, native applications, and the web from a single request. Generally available in the Mac app; rolling out to Pro and Max subscribers on Windows.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Research¶
Research: 17 tools · 1 Licensed
Literature search, evidence synthesis, and manuscript checking.
AI for Research maps these tools to each stage of a project, starting with Scopus with AI, licensed through the AUA Library; Gemini Notebook has its own guide because it works from your own uploads rather than the open web.
Asta¶
Status: Listed
Allen Institute for AIfree
Scholarly research assistant from the Allen Institute for AI, the makers of Semantic Scholar, that finds papers, answers questions with citations to the literature, and summarizes evidence. Free to use with a sign-in.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Claude Science¶
Status: Listed
Anthropicpaid
Local-first research workbench pairing an agent with curated scientific skills (genomics, proteomics, cheminformatics), local code execution, and a reviewer agent that checks citations and calculations.
Beta on paid Claude plans (on Team and Enterprise, an administrator turns it on first); for macOS, Windows, and Linux.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Consensus¶
Status: Listed
Consensusfreemium
Search engine that answers questions by summarizing findings across published, peer-reviewed studies.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Covidence¶
Status: Listed
Covidencepaid
Web platform for managing systematic reviews with a team, covering screening, data extraction, and reporting aligned to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) standard. Artificial intelligence features sort records by predicted relevance, tag randomized controlled trials, and suggest data extraction values that reviewers accept or reject one at a time.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Elicit¶
Status: Listed
Elicitfreemium
Literature research assistant that finds papers and extracts key findings into structured tables.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Gemini Notebook¶
Status: Listed
Googlefreemium
Research notebook that answers questions and generates summaries from documents you upload, with citations back to the passage; optional features can add web sources. Formerly NotebookLM.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Google Scholar Labs¶
Status: Listed
Googlefree
Artificial intelligence search mode in Google Scholar for detailed research questions. It identifies a question's key topics, searches Scholar, and lists papers that address it, each with a short note on how the paper helps, and it takes follow-up questions. Free with a Google account sign-in.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
OpenEvidence¶
Status: Listed
OpenEvidencefree
Clinical question-answering tool for verified health professionals, with answers cited to the medical literature.
Verification requires US credentials, either a National Provider Identifier (NPI) or enrollment at a US medical school; withdrew from the United Kingdom and European Union in spring 2026.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Prism¶
Status: Listed
OpenAIfree
Free online LaTeX editor and writing workspace for scientific papers, with ChatGPT built in for drafting, proofreading, citations, and literature search, and unlimited collaborators. Requires a personal ChatGPT account.
Checked October 2, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Rayyan¶
Status: Listed
Rayyan Systemsfreemium
Web platform for systematic and literature reviews that handles reference import, duplicate detection, and title and abstract screening across a review team. Artificial intelligence features rank records by predicted relevance on every plan including the free one, while automated data extraction requires the Advanced plan or higher.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
ResearchRabbit¶
Status: Listed
ResearchRabbitfreemium
Visual literature mapping from seed papers, surfacing related work, shared authorship, and citation neighborhoods. Free for searches from up to 50 seed papers, with a paid ResearchRabbit+ plan for larger searches and multiple projects.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
SciSpace¶
Status: Listed
SciSpacefreemium
Research workspace that combines literature search, question answering over uploaded papers, data extraction into tables, and citation formatting in one interface.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Scite¶
Status: Listed
Scite (Research Solutions)paid
Citation analysis tool that shows whether later papers support or dispute a given study's claims.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Scopus with AI¶
Status: Licensed
ElsevierAUA-licensed
Elsevier's Scopus citation database with its Scopus AI layer, answering plain-language questions with cited summaries, concept maps, and an agentic Deep Research mode, grounded in the abstracts of the indexed peer-reviewed literature.
Licensed institution-wide through the AUA Library (2026) and reviewed ahead of procurement. Sign in with AUA credentials via the library link.
What AUA provides: access, through a university license.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Semantic Scholar¶
Status: Listed
Allen Institute for AIfree
Free academic search engine with paper summaries, citation graphs, and author pages across scientific fields.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Undermind¶
Status: Listed
Undermindfreemium
Literature search assistant that runs a deep, iterative search on a topic you describe and returns a structured report with citations. Suited to thorough searches such as novelty checks rather than quick lookups.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
UpToDate Expert AI¶
Status: Listed
Wolters Kluwerpaid
Conversational tool inside UpToDate that answers clinical questions using only UpToDate's expert-written content, showing the assumptions and reasoning behind each answer, with links to the supporting UpToDate topics and drug information. Offered in select markets through participating institutional UpToDate subscriptions and individual subscriptions, so access through a school's UpToDate license depends on that license.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Medical Learning¶
Medical Learning: 12 tools
Practice questions, case work, and tutoring for medical study.
AMBOSS¶
Status: Listed
AMBOSSpaid
Medical knowledge platform and question bank with a built-in artificial intelligence study mode (AI Mode Learning) that answers questions grounded in the AMBOSS library, plus a beta collection of Assistants for case simulation, checklists, and consult preparation.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Aquifer¶
Status: Listed
Aquiferpaid
Library of more than 250 virtual patient cases across core clinical disciplines such as family medicine, internal medicine, and pediatrics. Artificial intelligence gives immediate feedback on the summary statements students write in the cases, and a growing set of Virtual Patient Encounters lets learners interview a simulated patient and get feedback on their history taking and diagnostic reasoning. Schools subscribe for their students, and individuals can buy a yearly plan that includes the summary statement feedback.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Medical Student AI¶
Status: Listed
Help You Match LLCfreemium
Tutoring app with a question bank in the style of the United States Medical Licensing Examination (USMLE), cited explanations, and custom quizzes built from uploaded study materials.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Nabu Tutor¶
Status: Listed
Nabu Tutorfreemium
Tutor for USMLE and Comprehensive Osteopathic Medical Licensing Examination (COMLEX) preparation that works through cases Socratically with follow-up questions and builds a study plan around weak areas.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Neural Consult¶
Status: Listed
Dendritic Health AIfreemium
Study platform for medical, nursing, and physician assistant students with flashcard generation from your own materials, board-style questions, and simulated patient encounters for clinical exam practice.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Oncourse AI¶
Status: Listed
Oncourse AIfreemium
Adaptive USMLE study platform whose lessons, flashcards, and quizzes adjust to performance, with a built-in tutoring chatbot.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Osmosis AI¶
Status: Listed
Elsevierpaid
Conversational study companion from Elsevier for medical students. Answers medical study questions with citations to the relevant section of an Elsevier textbook or Osmosis video, links answers to Osmosis videos, and generates flashcards and quiz questions on request.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
SimChat¶
Status: Listed
Geeky Medicsfreemium
Virtual patient platform from the Geeky Medics team in which educators turn their own patient scripts, objective structured clinical examination (OSCE) checklists, or case vignettes into artificial intelligence patients that learners interview by text, voice, or video avatar, with feedback against the educator's own marking criteria. The free plan covers up to five learners and two custom scenarios in text, with a one-time voice and video trial; ready-made case libraries, available on paid plans, are mapped to United Kingdom exams.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
StepGenie¶
Status: Listed
StepGeniefreemium
Question bank for USMLE Steps 1, 2 CK, and 3 whose answer explanations are built as sequences of images rather than text, alongside a 45-day study plan.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
StudyFetch¶
Status: Listed
StudyFetchfreemium
Study platform that turns uploaded course materials (slides, PDFs, recordings) into flashcards, quizzes, and practice tests, with an AI tutor that answers from what you upload. Free plan with limits; paid plans lift them.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Thea¶
Status: Listed
Theafreemium
Study app that builds flashcards, practice tests, and summaries with spaced repetition from uploaded readings or lecture notes.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
UAsk¶
Status: Listed
UWorldpaid
Artificial intelligence study assistant built into UWorld's question banks for the United States Medical Licensing Examination (USMLE) Steps 1 through 3. Answers questions about the item on screen, walks through the clinical reasoning, and can turn responses into flashcards and notebook entries.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Presentations and Design¶
Presentations and Design: 8 tools
Slide decks, posters, and visual design.
For a new deck, Claude Slides, a feature of Claude on every plan including Free, builds one from notes, a report, or a conversation and exports it to PowerPoint, PDF, or Google Slides. To improve an existing PowerPoint deck without rebuilding it, start with tools that work on the .pptx file itself: PowerPoint Design Suggestions applies suggestions inside the file, and Claude Design and Canva import a PowerPoint deck and export the result back to one. Gamma builds in its own format first and its PowerPoint export can shift layouts; Beautiful.ai also works in its own format but exports to PowerPoint cleanly.
Beautiful.ai¶
Status: Listed
Beautiful.aipaid
Presentation tool whose built-in design rules keep slides tidy as content changes. Works in its own format but exports to PowerPoint cleanly.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
BioRender¶
Status: Listed
BioRenderfreemium
Web-based tool for building scientific figures, slides, and posters from a library of biomedical icons and templates. Its artificial intelligence features convert uploaded images to the platform's style, edit icons from text prompts, and draft flowcharts, protocols, and timelines from written descriptions.
The free tier is for educational use only and grants no publication rights; check your plan before a figure goes into a manuscript.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Canva¶
Status: Listed
Canvafreemium
Design platform with prompt-based Magic Design features and a large asset library. Imports an existing PowerPoint deck for restyling and exports the result back to PowerPoint.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Claude Design¶
Status: Listed
Anthropicfreemium
Design feature in Claude, from Anthropic, that turns prompts and uploaded Word, PowerPoint, or Excel files into one-pagers, interactive prototypes, and other visual work inside a Claude conversation. Exports to PowerPoint, PDF, or web page (HTML) files, or sends work to tools such as Canva; for slide decks, Anthropic now points to Claude Slides, a separate feature of the same app.
On every Claude plan, including Free, in any Claude conversation and in Claude Code; the standalone claude.ai/design site closes on December 14, 2026. On Enterprise plans it turns on by default on October 15, 2026.
Checked October 9, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Gamma¶
Status: Listed
Gammafreemium
Tool that drafts presentations, documents, and simple web pages from a text prompt or outline.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Mentimeter¶
Status: Listed
Mentimeterfreemium
Interactive presentation tool for live polls, quizzes, word clouds, and anonymous questions that the audience answers from their own devices. Its artificial intelligence features draft a whole interactive session from a prompt or pasted outline, and turn a topic or pasted document into quiz questions.
Checked October 2, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Napkin AI¶
Status: Listed
Napkin AIfreemium
Generates editable diagrams, flowcharts, and infographics from pasted or imported text, a quick way to add visuals to lectures and handouts. The free plan exports images and PDFs with Napkin branding on the visuals; PowerPoint export requires a paid plan.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
PowerPoint Design Suggestions¶
Status: Listed
Microsoftinstitutional
Built into PowerPoint under the institutional Microsoft 365 license, where the feature, also called Designer, appears as Design Suggestions on the Design or Home tab. Suggests layouts, imagery, and design treatments for the slide being edited, so an existing deck improves one slide at a time without leaving the file.
What AUA provides: not confirmed. This tool needs an organization's license.
Checked September 29, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Image Generation¶
Image Generation: 4 tools
Still images from text prompts and references.
Adobe Firefly¶
Status: Listed
Adobefreemium
Adobe's generative AI app for images, video, and audio. Adobe's own Firefly models are trained on licensed and public-domain content, which matters when training-data provenance or commercial use is a concern; the app also offers partner models from companies such as Google, OpenAI, and Runway, whose suitability Adobe leaves to the user, so check which model is selected. Free daily generations; paid plans add monthly credits.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
ChatGPT Images¶
Status: Listed
OpenAIfreemium
Image generation and editing inside ChatGPT. Creates images from a description and edits existing ones, including adding text, changing details, and making a background transparent. Available on every tier, with slower and more limited generation on the free plan.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Ideogram¶
Status: Listed
Ideogramfreemium
Image generator designed for legible text inside images, useful for posters, badges, and labeled figures.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Midjourney¶
Status: Listed
Midjourneypaid
Image generation service used through its website or Discord, which can also turn an image into a 5-second video. No free tier.
Open by default, so every image and video can appear on the public Explore page on midjourney.com, including ones made in private Discord servers; Stealth Mode, which hides them, comes only with the Pro and Mega plans.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Video Generation¶
Video Generation: 6 tools
Short video clips from text or image prompts.
Flow¶
Status: Listed
Googlefreemium
Google's creative studio for video and image generation, combining the Veo 3.1 and Gemini Omni video models with Nano Banana image generation. It absorbed Google's earlier ImageFX and Whisk tools.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Hailuo¶
Status: Listed
MiniMaxfreemium
Fast video generator from the Chinese lab MiniMax with a usable free tier, a practical way to try video generation without paying.
Operated from Singapore; its privacy terms do not say where your uploads are stored for most users, which is worth knowing before you use it for work material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
HeyGen¶
Status: Listed
HeyGenfreemium
Generates presenter-style videos with artificial intelligence avatars, including translated, lip-synced versions of existing recordings.
A custom avatar or voice of someone other than you speaks in their name, so it is worth agreeing its use with them and labeling the result as generated.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Kling¶
Status: Listed
Kuaishoufreemium
Video generator from the Chinese company Kuaishou. The Kling 3.0 family generates clips with native synchronized audio and multi-language lip sync.
Operated from Singapore; its privacy terms say your data is stored on servers there and may be shared within its corporate group, which is worth knowing before you use it for work material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Runway¶
Status: Listed
Runwayfreemium
Video generation and editing platform aimed at production work, with fine control over camera moves, motion, and character consistency.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Synthesia¶
Status: Listed
Synthesiafreemium
Avatar video platform for training and instructional content. Turns a script or document into a narrated presenter video without filming.
A custom avatar or voice of someone other than you speaks in their name, so it is worth agreeing its use with them and labeling the result as generated.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Music and Audio¶
Music and Audio: 4 tools
Music, voice, and sound generation.
Adobe Podcast¶
Status: Listed
Adobefreemium
Web tool from Adobe for cleaning up and editing spoken audio. Its Enhance Speech filter removes background noise and echo from voice recordings, free with usage limits; the premium tier, included with the Adobe Firefly Pro plan, can also separate speech, background noise, and music into layers that are adjusted independently.
Checked October 12, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
ElevenLabs¶
Status: Listed
ElevenLabsfreemium
Speech synthesis, voice, music, and sound effect generation for narrating course material or producing audio versions of written content. Its music model was built in partnership with artists, labels, and publishers, and tracks are cleared for commercial use on paid plans, within limits that vary by plan (the free plan is for personal use).
A cloned voice of someone other than you speaks in their name, so it is worth agreeing its use with them and labeling the result as generated.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Suno¶
Status: Listed
Sunofreemium
Generates complete songs with vocals from a text description, with a free tier.
Its training data is the subject of ongoing copyright litigation, and its terms allow commercial use only of songs made on a paid plan and downloaded within that plan's monthly allowance, which matters before a track goes into published or shared material.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Udio¶
Status: Listed
Udiofreemium
Song generator that settled with Universal Music Group and Warner Music Group in 2025. Downloads of generated audio have been disabled since October 2025, pending its licensed platform, so tracks currently cannot leave the platform.
Its training data is the subject of ongoing copyright litigation, and its terms allow only personal, noncommercial use of generated songs, shared as links to the platform.
Checked September 23, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Writing¶
Writing: 3 tools
Drafting, editing, and grammar support.
Grammarly¶
Status: Listed
Superhuman (formerly Grammarly, Inc.)freemium
Writing assistant that checks grammar, clarity, and tone in documents, email, and browser text fields.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Paperpal¶
Status: Listed
Cactus Communicationsfreemium
Academic writing assistant with grammar and language checking tuned to scholarly text, plus paraphrasing, plagiarism checks, and journal submission readiness checks. Works in Microsoft Word, Google Docs, Overleaf, and a web editor.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Writefull¶
Status: Listed
Digital Sciencefreemium
Language feedback for research manuscripts from models trained on published journal articles, used through Microsoft Word and Overleaf add-ins, with a revision tool for checking a full document before submission.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Meetings and Transcription¶
Meetings and Transcription: 4 tools
Meeting capture, transcription, and notes.
Fathom¶
Status: Listed
Fathomfreemium
Meeting notetaker that joins video calls to record, transcribe, and summarize them, delivering the summary shortly after the call ends. The free plan does not cap recordings or transcription.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Microsoft Teams Transcription¶
Status: Listed
Microsoftinstitutional
Built-in live transcription for Teams meetings under the institutional Microsoft 365 license.
What AUA provides: not confirmed. This tool needs an organization's license.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Otter¶
Status: Listed
Otter.aifreemium
Meeting assistant that transcribes and summarizes video calls, with live transcripts that participants can highlight and comment on during the meeting. The free plan includes 300 transcription minutes per month.
Checked September 15, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Whisper¶
Status: Listed
OpenAIfree
Open-source speech-to-text model that transcribes audio recordings on your own computer.
Runs locally; audio never leaves your computer.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Local Models¶
Local Models: 4 tools
Apps for running open-weights models on your own machine.
Running Models Locally is the walkthrough for these apps, and Hardware for Local AI sizes the machine.
Jan¶
Status: Listed
Jan (open source)free
Open-source desktop application for downloading and chatting with open-weights models fully offline, similar in role to LM Studio but Apache 2.0 licensed. Runs on Windows, macOS, and Linux.
Checked September 22, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
LM Studio¶
Status: Listed
LM Studiofreemium
Desktop application with a graphical interface for downloading and chatting with local language models, which remains a free download. The vendor's lead product is now LM Studio Bionic, an agent for open models with free local use and optional paid cloud models.
Checked October 6, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Ollama¶
Status: Listed
Ollamafree
Tool for downloading and running open-weights models on your own computer, through a command line or an official desktop chat app. The local app is free; the company separately sells an optional cloud service for models too large to run locally.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
Open WebUI¶
Status: Listed
Open WebUIfree
Self-hosted web interface for chatting with local models served by tools such as Ollama.
Checked October 13, 2026: the date this listing's details (link, description, and cost) were last checked; not a review of the tool
For a whole task walked through end to end, including how to check the result, the step-by-step guides take one at a time, and the Prompt Library holds the templates they draw on.
Open-Weights Models¶
The assistants above run in their vendor's cloud. Open-weights models are different: the model file itself is published for anyone to download and run on their own hardware, so nothing you type leaves your machine. The Local Models tools above are the usual way to run them; Running Models Locally is a practical walkthrough. Capability rankings shift quickly; the Benchmarks page tracks current standings. Model families carry a license instead of a governance status, because there is no service operator to have a relationship with. The license named on each entry says what you may do with a model, such as commercial use or redistribution.
Open-weights model families and notable single releases; each entry shows the date its license and description were last checked:
Language¶
Language: 21 model families
Chat, reasoning, and coding model families, from laptop-sized models to data-center scale.
PrismMLApache 2.0
PrismML's Ternary Bonsai 2, a 27-billion-parameter model derived from Qwen3.8 27B and compressed to about 6 gigabytes, small enough for a standard laptop, with a 262,000-token context window. It needs PrismML's own builds of llama.cpp or MLX rather than the standard local apps.
Checked September 22, 2026
DeepSeekMIT
DeepSeek publishes the weights of its flagship chat and reasoning models, including the V4 series, under the permissive MIT license.
Checked June 10, 2026
Domyn (formerly iGenius)MIT-style (text on domyn.com)
European 10 billion parameter reasoning model from the Milan company Domyn, with a dual thinking mode toggleable from the system prompt, a 32K native context extensible to 128K, and coverage optimized for English, Italian, Spanish, French, and German. Its MIT-style license permits commercial use.
Checked July 28, 2026
GoogleApache 2.0
Google's open model family, released April 2026, from 2 billion parameter on-device variants to a 31 billion parameter model, with vision and audio input and support for over 140 languages.
Checked June 10, 2026
Z.ai (Zhipu AI)GLM-5.3 License
The current flagship of the GLM open-weight family behind the Z.ai assistant, a mixture-of-experts model that keeps the 5.2 base model and gains its improvements from post-training, with the largest gains in complex coding. Supported by local toolchains including llama.cpp. Published under a custom license that permits use, modification, and redistribution, with conditions that apply only to very large commercial providers.
Checked September 1, 2026
OpenAIApache 2.0
OpenAI's open-weight pair. gpt-oss-120b runs on a single 80 gigabyte graphics card with reasoning performance near the company's o4-mini, and gpt-oss-20b runs on machines with 16 gigabytes of memory.
Checked June 10, 2026
IBMApache 2.0
IBM's open-weight family, built for tool calling and agent work at sizes that run on ordinary hardware. The 4.2 release comes in 3 billion, 8 billion, and 30 billion parameter versions under Apache 2.0, with quantized builds published alongside for lower-memory machines.
Checked September 1, 2026
TencentApache 2.0
A very large open-weight mixture-of-experts model from Chinese company Tencent, 770 billion total parameters with 49 billion active per token and a 1 million token context window, under Apache 2.0. Data-center hardware only. It succeeds Hy3, which was roughly a third the size.
Checked September 1, 2026
Thinking Machines LabApache 2.0
Thinking Machines Lab's first open-weight model, a mixture-of-experts design with 975 billion total parameters (41 billion active per token) and a context window of up to 1 million tokens, under Apache 2.0. Data-center hardware only.
Checked July 21, 2026
Moonshot AIModified MIT
Moonshot AI's open-weight family, aimed at coding and multi-step agent tasks. The flagship K3, at 2.8 trillion total parameters with a 1 million token context window, now has published weights under an MIT-style license whose extra conditions apply only to very large commercial deployments; a model that size needs data-center hardware, not a consumer graphics card.
Checked August 25, 2026
Aleph AlphaApache 2.0
The German company Aleph Alpha's open-weight reasoning model for German and English. Kolibri 1, released in October 2026, is a mixture-of-experts model with 78 billion total parameters and about 3.5 billion active per token and a context window of up to about 1 million tokens, under Apache 2.0. Its roughly 78 gigabytes of weights need data-center graphics cards.
Checked October 12, 2026
MetaLlama Community License
Meta's open-weight family. Llama 4 Scout and Maverick handle text and images natively, and Scout offers a context window of up to 10 million tokens.
Checked June 10, 2026
MeituanMIT
A very large open-weight mixture-of-experts model from Chinese company Meituan, 1.6 trillion total parameters with about 48 billion active per token and a 1 million token context window, under MIT.
Checked July 2, 2026
GoogleHealth AI Developer Foundations
Google's open medical variants of Gemma, trained for medical text and image comprehension across radiology, pathology, dermatology, and medical records. Downloading requires accepting Google's Health AI Developer Foundations terms, which permit research and commercial use. Google presents it as a starting point for developers, and its model card says outputs are not intended to directly inform clinical diagnosis or treatment without validation.
Checked July 15, 2026
MiniMaxMiniMax Community
Large open-weight mixture-of-experts model from the Chinese company MiniMax, with about 428 billion total parameters and 23 billion active per token. Accepts text, image, and video input with a 1 million token context window.
Checked July 15, 2026
Mistral AIApache 2.0
French lab Mistral's open-weight range, from the small Ministral models to Mistral Large 3, with most flagship releases under Apache 2.0.
Checked June 10, 2026
MetaApache 2.0
Meta's 30 billion parameter open-weight model, accepting image and text input and optimized for agentic task completion and local use, with official GGUF and on-device variants. Notably Meta's first release under plain Apache 2.0 rather than the Llama community license.
Checked August 11, 2026
NVIDIAOpenMDW 1.1
NVIDIA's open-weight mixture-of-experts model, 30 billion total parameters with about 3 billion active per token, sized for local and workstation use. Released under the Linux Foundation's permissive OpenMDW license rather than an NVIDIA-specific one.
Checked August 18, 2026
Allen Institute for AI (Ai2)Apache 2.0
A fully open language model family from the Allen Institute for AI (Ai2), a Seattle nonprofit, which publishes the training data, training code, and intermediate checkpoints alongside the weights, so researchers can study how the model was built and trace its output back to the training data. Released in 7 billion and 32 billion parameter versions, each with a chat (Instruct) and a step-by-step reasoning (Think) variant, under Apache 2.0.
Checked September 29, 2026
HuaweiOpenPangu license
Huawei's open-weight mixture-of-experts model, 92 billion total parameters with 6 billion active and a 512k context window, trained on Huawei's own accelerators under a custom license.
Checked July 2, 2026
AlibabaApache 2.0 (Qwen3.8-Max license on the 2.4T flagship; Qwen Community License 1.0 on Flash-Next)
Alibaba's open-weight line, offered in a wide range of sizes, from models small enough for a laptop to flagship scale, with broad multilingual coverage. The August 2026 Qwen3.8 release opened a Max-class flagship for the first time (2.4 trillion parameters with 95 billion active, data-center hardware only) alongside a 27 billion parameter vision-language model that runs on a workstation, both with a 262,000 token context extensible to about 1 million. Qwen3.8-Flash-Next, published later that month as a preview of the next architecture, is multimodal and carries a third license.
Checked September 1, 2026
Image Generation¶
Image Generation: 4 model families
Image generators for local or self-hosted pipelines.
Black Forest LabsFLUX Non-Commercial (dev)
Black Forest Labs' open-weight image family. FLUX.2 dev (32 billion parameters) generates and edits images under a non-commercial license, with a smaller klein variant for lighter hardware. Runs locally in ComfyUI.
Checked July 15, 2026
AlibabaApache 2.0
Image generation and editing model under a fully permissive license, notably strong at rendering legible text inside images.
Checked June 12, 2026
Stability AICommunity License
A widely supported image generation line with a large ecosystem of community fine-tunes and tooling. Free for individuals and small organizations under Stability's community license.
Checked June 12, 2026
AlibabaApache 2.0
Alibaba's 6 billion parameter image generation family, sized to run on a computer with a 16 gigabyte graphics card. The fast Turbo version renders English and Chinese text inside images, and the base version is published for fine-tuning, both under a fully permissive license.
Checked October 2, 2026
Video Generation¶
Video Generation: 3 model families
Video generators for local or self-hosted pipelines, generally needing more video memory than image models.
TencentCommunity License
Tencent's open-weight video model, designed for natural motion and physics such as fluids, cloth, and object interactions.
Checked June 12, 2026
LightricksLTX-2 Community License
Lightricks' open-weight family that generates video with synchronized sound in a single model; the version linked, LTX-2.3, has 22 billion parameters and runs locally in ComfyUI. Free under its community license for organizations with annual revenue under US$10 million, counted together with affiliated companies; larger organizations need a paid license from Lightricks, and the license makes no exception for academic use. The newer LTX-2.5 uses a separate license with the same revenue threshold and requires signing in to Hugging Face to download.
Checked September 29, 2026
AlibabaApache 2.0
Open-weight video family focused on photorealism, especially for human subjects. Top quality wants a 24 gigabyte graphics card.
Checked June 12, 2026
Music and Audio¶
Music and Audio: 5 model families
Speech, music, and sound models.
ACE Studio and StepFunMIT
Open-weight song generation model producing full tracks with vocals, from short loops to compositions of up to 10 minutes, and following prompts in more than 50 languages. Version 1.5 runs on a graphics card with less than 4 gigabytes of memory, under the MIT license; its developers say it was trained on licensed, royalty-free, and synthetic music.
Checked September 29, 2026
hexgrad (independent developer)Apache 2.0
A small text-to-speech model, 82 million parameters, that reads text aloud in 54 preset voices across eight languages and is small enough to run on a personal computer, under Apache 2.0. More than half of its voices are American or British English, and its voice list notes that support for other languages may be thin. Version 1.0, from January 2025, is the current release; its developer says it was trained only on public domain, permissively licensed, or synthetic audio.
Checked September 29, 2026
GoogleHealth AI Developer Foundations
Google's open-weight speech recognition model for medical dictation, which turns recorded speech into text. It has 105 million parameters and was trained on about 5,000 hours of de-identified physician dictation in radiology, internal medicine, family medicine, and other specialties. English only, and most training speakers were native English speakers raised in the United States, so accuracy may be lower for other accents. Downloading requires accepting Google's Health AI Developer Foundations terms, and its model card says outputs may contain transcription errors and are not intended to directly inform clinical decisions.
Checked September 29, 2026
MetaCC BY-NC 4.0
Meta's research model for generating instrumental music from text descriptions, licensed for non-commercial use.
Checked June 12, 2026
Stability AIStable Audio Community
Generates music, sound effects, and production elements from text, with structured compositions up to several minutes. The small and medium weights are open on Hugging Face; the largest model is not released.
Checked July 15, 2026
Tabular Data¶
Tabular Data: 1 model family
Models for tabular and structured data.
Google ResearchTabFM Non-Commercial v1.0
Google Research's open-weight foundation model for tabular data, making zero-shot predictions on spreadsheet-style datasets without task-specific training. Relevant to research data analysis; weights are licensed for non-commercial use.
Checked July 7, 2026
What the Statuses Mean, and How to Request a Review¶
Each status describes where the institution stands with a tool, not a rating of its quality. Anyone at AUA can ask the AI Committee to review a tool through How Tools Are Reviewed.
| Status | What it means |
|---|---|
| Listed | In the directory because it is relevant and live. Not an endorsement; use your own judgment. |
| Licensed | Institutionally licensed or procured by the university. |
| Reviewed | Examined by the AI Committee; conclusions in the entry's note. |
| Use with caution | A specific concern, such as data handling, a consent requirement, or unresolved legal terms, is recorded in the entry's note. |
| Restricted | Reviewed and found unsuitable for institutional use. |
Where an entry has something specific worth knowing, such as an access condition, a consent requirement, or where the service is operated, it is printed on the card under the description. The word after each vendor's name shows cost: free, freemium (a free tier with paid upgrades), paid, institutional (needs an organization's license), or AUA-licensed (provided by AUA). The "Where to start" notes above a task's tools are suggestions drawn from each tool's own description and this site's guides; they are not statuses.
What AUA provides. Two tools are licensed by the university: Microsoft 365 Copilot and Scopus with AI. Two more need an organization's license, and this directory does not confirm AUA access to them: Microsoft Teams Transcription and PowerPoint Design Suggestions.
A Listed or Licensed status describes AUA's relationship with a tool, not an approval for sensitive data; the AI Responsible Use Policy is the reference for patient information, student records, and other sensitive material.
To suggest a tool, use the contact details on the About page. To request committee review, follow How Tools Are Reviewed. Reviewed, Use with caution, and Restricted come from that review; Listed and Licensed reflect cataloging and procurement facts.
Content last reviewed September 1, 2026. Review dates are maintained as described on the About page.