Skip to content

why cot

Why cot?

cot is a self-hosted observability layer for AI coding agents. It captures what your agents do, without changing how you work.

Why "cot"?

Short for chain of thought: the visible trail of prompts, tool calls, and outputs that make up an agent session. cot turns that trail into something you can browse, search, and understand.

Use cases

Debug agent runs

See exactly which tools fired, what files were touched, and where a session went sideways.

Understand usage

Track which agents, models, tools, tokens, and estimated spend you actually use across projects.

Review with context

Share session timelines with teammates instead of scrolling through terminal output.

How it fits on your machine

  1. 1

    YOUR AGENT

    Claude Code · Cursor · Codex

  2. 2

    LOCAL BRIDGE

    ~/.cot/bin/cot

  3. 3

    COLLECTOR

    FastAPI in Docker

  4. 4

    STORAGE

    ~/.cot/cot.db

  5. 5

    DASHBOARD

    localhost:31337

Recorded on your machine

  • Prompts and agent responses
  • Tool calls: shell, file edits, MCP, web search
  • Session metadata: project path, model, timestamps
  • Token counts (Claude sessions)

Never sent anywhere by default

  • No cloud accounts or logins
  • No prompts or responses sent to cot.run by default
  • No raw image bytes, attachment metadata only

get started

Start with your own agents.

One command. Self-hosted. Free and open source. See pricing

Building with agents across a team?
Talk to us about cot for organizations