agentleFS
snapsynapsesnapsynapse/ai-tool-watch/docs/llms.txt

ai-tool-watch

llms.txt11 starsChanged 4 months agoMITView on GitHub

What can this AI actually do? A plain-English reference for AI capabilities, plans, constraints, and implementations across hosted consumer AI products from providers with roughly 1% or greater public market share (ChatGPT, Claude, Gemini, Copilot, Grok, Perplexity, Meta AI). Maintained by Snap Synapse. Every claim carries its primary source and last-verified date. Self-hosted runtimes (3) and open-model access records (11) were frozen on 2026-10-11 when scope narrowed to hosted products and maintenance was slimmed. They keep their last-verified dates and…

The file

We split this llms.txt into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Start with the index: https://aitool.watch/api/v1/index.json

    line 29
  2. Very specific:

    Every record includes verified and checked dates for freshness

    line 47
  3. Very specific:

    Always include the verified date when citing data — the AI landscape changes weekly

    line 53
  4. Distinctive:

    Don't strip caveats from talking points — they contain important context about restrictions

    line 55
  5. Distinctive:

    All implementations of a capability across products.

    line 87

≈ “Common” is an estimate: at least 1 in 5 of the 3,592 llms.txt files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
llms.txt
# AI Tool Watch

> What can this AI actually do?

A plain-English reference for AI capabilities, plans, constraints, and implementations across hosted consumer AI products from providers with roughly 1% or greater public market share (ChatGPT, Claude, Gemini, Copilot, Grok, Perplexity, Meta AI).

Maintained by Snap Synapse. Every claim carries its primary source and last-verified date.

## What this site covers

- 18 capabilities (e.g., Search the Web, Generate Images, Write and Edit Code)
- 87 implementations across 7 subscription AI products
- Plan-level availability (which subscription tier unlocks what)
- Surface availability (web, desktop, mobile, terminal, API)
- Gating (free, paid, limited)
- Evidence with sources and changelog for every claim

## Archived (frozen 2026-10-11, not refreshed)

Self-hosted runtimes (3) and open-model access records (11) were frozen on 2026-10-11 when scope narrowed to hosted products and maintenance was slimmed. They keep their last-verified dates and may be out of date.

- [Archive page](https://aitool.watch/archive.html)
- [archive.json](https://aitool.watch/api/v1/archive.json)

## Machine-readable API

Stable JSON exports at: https://aitool.watch/api/v1/

Start with the index: https://aitool.watch/api/v1/index.json

### Available files

- [capabilities.json](https://aitool.watch/api/v1/capabilities.json): All capabilities with search terms, definitions, and cross-links
- [products.json](https://aitool.watch/api/v1/products.json): All products with provider links and implementation lists
- [implementations.json](https://aitool.watch/api/v1/implementations.json): All implementations with capability mappings and evidence
- [providers.json](https://aitool.watch/api/v1/providers.json): All providers with product lists
- [model-access.json](https://aitool.watch/api/v1/model-access.json): Archived open-model access records, frozen 2026-10-11
- [archive.json](https://aitool.watch/api/v1/archive.json): All records frozen 2026-10-11 (self-hosted runtimes, open-model access, providers)
- [evidence.json](https://aitool.watch/api/v1/evidence.json): Evidence records with sources and changelog
- [capability-matrix.json](https://aitool.watch/api/v1/capability-matrix.json): Capability x product availability grid
- [product-comparisons.json](https://aitool.watch/api/v1/product-comparisons.json): Pairwise product capability overlap
- [plan-entitlements.json](https://aitool.watch/api/v1/plan-entitlements.json): What each plan tier unlocks per product

### Data contract

- IDs are stable and safe to use as foreign keys
- Every record includes `verified` and `checked` dates for freshness
- File paths under /api/v1/ are stable within the version
- New fields may be added; existing fields will not be removed within v1

### Usage guidance

- Always include the `verified` date when citing data — the AI landscape changes weekly
- Respect gating and constraints — "available" does not mean "available to everyone"
- Don't strip caveats from talking points — they contain important context about restrictions
- Cache for up to 24 hours; don't poll more than once per hour
- Attribution appreciated: "Data from AI Tool Watch (https://aitool.watch/)"

## Agent access

- [agents.json](https://aitool.watch/agents.json): Agent capability manifest describing API endpoints and MCP tools
- [MCP server](https://github.com/snapsynapse/ai-tool-watch/blob/main/scripts/mcp-server.js): 15 read-only tools over stdio JSON-RPC; speaks MCP spec 2026-07-28 (stateless, server/discover) and legacy 2024-11-05
- [RSS feed](https://aitool.watch/index.xml): Recent feature changes as RSS 2.0

## Site pages

- [Capabilities](https://aitool.watch/): Capability-first homepage (what can AI do?)
- [Products](https://aitool.watch/implementations.html): Product-first detailed availability
- [Compare](https://aitool.watch/compare.html): Side-by-side product comparison
- [Limits](https://aitool.watch/constraints.html): Access tiers, surfaces, and limits
- [Timeline](https://aitool.watch/timeline.html): Chronological view of feature launches and changes
- [About](https://aitool.watch/about.html): About the project
- [Pattern](https://aitool.watch/pattern.html): Knowledge as Code: the development pattern behind this project
- [Archive](https://aitool.watch/archive.html): Self-hosted runtimes and open-model access, frozen 2026-10-11

## Bridge pages (programmatic)

### Capability checks: /can/{product}/{capability}/
Answers "Can {product} {capability}?" with plan tables, constraints, and evidence.
Example: https://aitool.watch/can/chatgpt/search-the-web/

### Product comparisons: /compare/{product-a}-vs-{product-b}/
Side-by-side capability overlap between two products.
Example: https://aitool.watch/compare/chatgpt-vs-claude/

### Capability landings: /capability/{slug}/
All implementations of a capability across products.
Example: https://aitool.watch/capability/search-the-web/

### Use-case pages: /best-for/{use-case}/
Product coverage grid for a use-case category.
Example: https://aitool.watch/best-for/creative-work/

## Source

- [Repository](https://github.com/snapsynapse/ai-tool-watch): GitHub
- License: MIT
- [Contact](https://github.com/snapsynapse/ai-tool-watch/issues): GitHub Issues

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.