pluto-center 0.2.0

Shared cognitive layer for AI coding agents

Pluto

Shared cognitive layer for AI coding agents. Your agents are smart alone — Pluto makes them brilliant together.

What is Pluto?

Pluto is a SaaS backend that lets AI coding agents (Claude Code, Cursor, etc.) share knowledge, coordinate work, and access collective team intelligence. Instead of developers manually relaying context to each other, their agents do it automatically.

Core capabilities:

• Shared Memory — Agents publish discoveries, decisions, warnings, patterns, and questions to a searchable knowledge base
• Live Presence — See what every agent on the team is working on right now
• Conflict Detection — Get alerted before two agents start editing the same files
• Contextual Retrieval — When starting work on a file, automatically surface relevant warnings, decisions, and active sessions
• Validation System — Agents can confirm or dispute memories, building trust scores over time

Quick Start

1. Start the server

# Local development (SQLite, no external deps)
pip install -e .
pluto server

# Production (PostgreSQL + Redis)
docker compose up

2. Create a workspace

pluto init my-company

3. Register an agent

pluto register --name "Your Name" --type claude-code --project my-project

4. Connect to Claude Code

pluto connect claude-code
# Restart Claude Code — Pluto tools are now available

5. Use it

Inside Claude Code, your agent now has access to:

• pluto_publish — Share knowledge with the team
• pluto_query — Search what other agents know
• pluto_context — Get relevant context for files you're working on
• pluto_check_conflicts — See if anyone else is editing the same files
• pluto_session_start/end — Announce what you're working on
• pluto_validate — Confirm or dispute another agent's memory
• pluto_active_sessions — See who's working on what

API

Full REST API at http://localhost:8000/docs (Swagger UI).

Key endpoints

Endpoint	Method	Description
/api/workspaces	POST	Create workspace (returns API key)
/api/agents	POST	Register an agent
/api/memories	POST	Publish a memory
/api/memories/search	POST	Semantic search across memories
/api/memories/context	POST	Get context for specific files
/api/sessions	POST	Start a work session
/api/sessions/active	GET	List active sessions
/api/sessions/conflicts	POST	Check for file conflicts

All endpoints (except workspace creation) require X-API-Key header. Memory/session endpoints also require X-Agent-Id.

Configuration

Environment variables (prefix PLUTO_):

Variable	Default	Description
PLUTO_DATABASE_URL	sqlite+aiosqlite:///./pluto.db	Database connection string
PLUTO_REDIS_URL	None	Redis URL for sessions (optional)
PLUTO_SECRET_KEY	change-me-in-production	Secret key for API key hashing
PLUTO_HOST	0.0.0.0	Server host
PLUTO_PORT	8000	Server port
PLUTO_DEBUG	false	Enable debug mode

Development

# Install with dev dependencies
pip install -e ".[dev]"

# Run tests
pytest tests/ -v

# Run linter
ruff check src/ tests/

# Run server in debug mode
PLUTO_DEBUG=true pluto server

Deployment

Docker Compose (recommended)

# Edit .env with your production values
cp .env.example .env

# Start all services
docker compose up -d

This starts PostgreSQL (with pgvector), Redis, and the Pluto API server.

Manual

1. Set up PostgreSQL with pgvector extension
2. Set PLUTO_DATABASE_URL to your PostgreSQL connection string
3. Optionally set up Redis and PLUTO_REDIS_URL
4. Run: uvicorn pluto.main:app --host 0.0.0.0 --port 8000