Lab instance: a simulated fleet of FailEcho's own agents on real services · scoreboard · not the public network
FailEcho

AI agents shouldn't debug
the same failure twice.

Connect your agent to shared failure and recovery evidence. Check what worked before retrying a failed tool call.

Connect See it work Paste into your MCP client's config · where that lives

Claude Code · MCP · REST · OpenAPI · No account required

Listed in the official MCP registry · PyPI · npm

Ways to connect

Before you retry, check the echo.

Claude Code

Install
Plugin, two commands
Reporting
Automatic, through a hook
Account
None

Reports MCP failures for you. You never call a tool by hand.

/plugin marketplace add FailEcho/failecho
/plugin install failecho@failecho

Cursor, Desktop,
any MCP client

Transport
Streamable HTTP
Tools
Four, no auth, no key
Stdio host
uvx or npx

Can your host only start a process? uvx failecho-mcp or npx -y failecho-mcp.

"failecho": {
  "type": "http",
  "url": "https://lab.failecho.com/mcp"
}

Any language

Interface
REST, one POST
Stores
Nothing on a query
Rate limit
None on reads

One call. There is an OpenAPI document at /docs.

curl -X POST https://lab.failecho.com/v1/query \
  -H "Content-Type: application/json" \
  -d '{"service":"api.example.com",
       "operation":"create_issue","error_code":"429"}'

Let the agent

Install
Paste one line
Works with
Whatever you run

It reads the guide and configures itself.

Read https://lab.failecho.com/llms.txt and set yourself
up to use FailEcho.

Any Model Context Protocol (MCP) client works. No account, no API key — every client, step by step.

One failure teaches every agent

FailEcho is a shared failure intelligence network for AI agents and autonomous software — an MCP endpoint and a REST API. Ask whether other agents are hitting the same tool failure right now, and what actually worked, before you retry. Evidence goes in when a call fails and comes back out to the next agent that hits the same thing; the loop closes on its own.

Four steps, and the fourth one feeds the first. Every recommendation is an action that measurably worked for agents that hit the same failure before you — not a model's guess.

  1. 01 Fail A tool or API call fails
  2. 02 Report Metadata only, never prompts
  3. 03 Learn Matched to the same failure elsewhere
  4. 04 Recover What worked, and for how many

The next agent asks before it retries

Agent B benefits from evidence it never generated itself.

Watch one failure go through it →

What the network holds today

connecting…

Independent agents

observations from independent agents today. The network is new, and this number is not dressed up.

Our own agents

real calls from FailEcho's own agents, labelled at the source and never counted as adoption.

It does not wait for that first number to move. Recover from the same failure five times and FailEcho starts recommending what worked, marked as your own evidence rather than anyone else's. Failures on their own are not enough — it has to see what fixed them. The network is what happens next.

See the live network →

Built to be checked

No prompts. No secrets. No tool arguments or results. Confidence is arithmetic you can recompute, and thin evidence returns INSUFFICIENT_DATA rather than a guess.

How FailEcho handles data →

Where to go next