apifacts
Registry of verified API behaviour that agents query when an HTTP call fails
- Pricing
- Open source
- Platforms
- CLI, API
- Runs
- Locally
- Version
- 0.2.1
- Listing updated
- 2 days ago
About apifacts
apifacts is an open-source MCP server for developer tools. It runs locally on your machine and is started by your AI client.
Docs drift, and an agent that hits a 200 with an empty array instead of a 429 wastes a session rediscovering the trap. apifacts keeps dated, typed facts about how third-party HTTP APIs, SDK packages and CLIs behave today, and answers with the matching fact and a fix.
The MCP server has a single tool, check, which takes a target, method, status and response body. Facts live as one YAML file each and are confirmed by scheduled, deterministic HTTP probes; the README says no language model is used in verification. Fixes come from a closed vocabulary with typed parameters rather than free text. An offline mode answers from a snapshot packaged in the wheel and sends nothing.
The same registry also ships as a pytest plugin, a CLI, a log scanner and a GitHub Action, so one fact base serves tests, CI and agents.
Install
Run in a terminal
claude mcp add apifacts -- uvx apifactsAdd to claude_desktop_config.json
{
"mcpServers": {
"apifacts": {
"command": "uvx",
"args": [
"apifacts"
]
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"apifacts": {
"command": "uvx",
"args": [
"apifacts"
]
}
}
}Add to .vscode/mcp.json
{
"servers": {
"apifacts": {
"type": "stdio",
"command": "uvx",
"args": [
"apifacts"
]
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"apifacts": {
"command": "uvx",
"args": [
"apifacts"
]
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"apifacts": {
"command": "uvx",
"args": [
"apifacts"
]
}
}
}Tools this server provides (1)
Taken from the project documentation; the running server may differ.
- check
- Takes target, method, status and body of a failing API call and returns matching verified facts with typed fixes.
What it does
- Single check tool
- One tool takes the target, method, status and body of a failed call and returns matching facts.
- Probe-verified facts
- Each fact is established by scheduled declarative HTTP probes, not by a language model.
- Typed fixes
- Suggested fixes are drawn from a closed vocabulary with typed parameters.
- Offline mode
- apifacts mcp --offline answers from the snapshot in the wheel and makes no requests.
- CI and test integrations
- A pytest plugin, a GitHub Action and a log scanner surface the same facts.
- Never an empty answer
- Responses carry facts, observations and an unknown reason when nothing matches.
Good for
- Look up why a Stripe or GitHub call returned an unexpected status
- Show known facts when a pytest test fails on an HTTP error
- Annotate a red CI build with verified API behaviour
- Let a coding agent check an API quirk before changing code
Pricing
Open source
Prices change. Check apifacts.dev before you buy.
apifacts FAQ
How do I add it to Claude Code?
Run claude mcp add apifacts -- uvx --from 'apifacts[mcp]' apifacts mcp.
Does it work without network access?
Yes. The --offline flag answers from the snapshot bundled in the wheel.
Is a model used to verify facts?
No. The README says verification relies only on deterministic probes.
What does the server expose?
One tool, check, and no resources or prompts.
Is apifacts free?
Yes. apifacts is open source, so you can use and self-host it for free.
How do I add apifacts to Claude Code?
Run this command in your terminal: claude mcp add apifacts -- uvx apifacts
What are the best alternatives to apifacts?
The closest alternatives to apifacts are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to apifacts

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools