# Observal/Observal

Observal is self-hosted registry for your coding agent extensions with a built in insight engine.  Setup Observal, define the scope and share your Skills, MCPs and Agents with your peers.

Repository: https://github.com/Observal/Observal
Canonical: https://ross.abutalabs.com/products/observal
Homepage: https://observal.io/
Language: Python
License: Apache-2.0
License Family: permissive
Topics: cli-tool, large-language-models, open-source, playground, self-hosted, agents, claude-code, cursor, kiro, antigravity, cursor-ai, insights, analytics, codex, mcp, registry, skills, pi
Last push: 2026-08-26T02:22:02+00:00

## Health v2 (maintenance only)
Score: 80/100 (v2, computed 2026-09-03T02:39:23.370411+00:00)
- activity 99, release rhythm 96, longevity 11
- inputs: {"age_days": 156, "days_push": 8, "days_rel": 24, "gap_med": 1, "n_releases_24m": 42}
- flags: young
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 2355, forks 472 (observed 2026-08-28T04:06:40.472393+00:00)

## What it is
Observal is a self-hosted registry and control plane for coding agent extensions such as Skills, MCP servers, and Agents, with a built-in insight engine for observing agent traffic and configs. It consists of a Docker Compose server stack (API, web UI, PostgreSQL, ClickHouse, Redis) and a per-developer CLI.

## Use cases
- run a team-wide registry of MCP servers and agent skills
- observe and debug MCP traffic from coding agents
- share agent configurations across harnesses like Claude Code and Cursor
- self-host a private registry for AI coding agent extensions
- track which agent components are outdated across a team
- debug agent failures with session tracking and reconciliation

## When to choose
- your team shares MCP servers, skills, or agent configs and needs a single source of record
- you want self-hosted control over agent component distribution and usage insights
- you use multiple coding agent harnesses (Claude Code, Cursor, Kiro, Copilot) and need consistent configs

## When to avoid
- you are a solo developer with no need to share agent components across a team
- you only need a public MCP marketplace rather than a private self-hosted registry
- you cannot operate a Docker Compose stack with PostgreSQL, ClickHouse, and Redis

## Facets
- artifact type: service
- maturity: active
- function: mcp, agent-framework, analytics, developer-tools, cli, self-hosted
- domain: developer-tools, large-language-models, self-hosted, analytics
- platform: self-hosted, cli, python
- tags: agent-registry, mcp-registry, skills-registry, claude-code, cursor, insights-engine, control-plane, teamspaces, ai-agents, docker, web-server

## Member repositories
- Observal/Observal (main) score 80

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:06:40.472393+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-30T02:36:26.133751+00:00, confidence not recorded.
  - readme: https://github.com/Observal/Observal (fetched 2026-08-28T04:06:40.472393+00:00, sha e14e7f880d79)
  - homepage: https://observal.io/ (fetched 2026-08-29T10:17:05.502392+00:00, sha 44136fa355b3)
  - site_page: https://docs.observal.io (fetched 2026-08-29T10:17:05.505412+00:00, sha f62e88912cdb)
- Data as of 2026-08-30T08:39:29.467469+00:00.