Documentation

Run Pulsyr and connect your first agent.

The quickest path uses Docker Compose and the published container image. You need Docker, a generated application secret, and a database password.

Before production: choose unique secrets, use HTTPS, keep DEBUG=false, and arrange database backups.

1. Get the Compose project

git clone https://github.com/rlcuevas14/PULSYR
cd PULSYR
cp .env.example .env

The Compose file starts Pulsyr and Postgres with pgvector support. It can build locally or use the image configured for your repository namespace.

2. Configure the required secrets

SECRET_KEY=<a random value of at least 32 characters>
DB_PASSWORD=<a unique database password>
DEBUG=false

Generate the application secret with a cryptographically secure tool. Pulsyr refuses to start in production with a distributed placeholder or short key.

3. Start the services

docker compose up -d

Open the configured host. A new database redirects to /setup, where you create the owner account, first project, and first write token in one flow. The raw token is shown once—store it in your MCP client's secret configuration.

4. Connect an MCP-compatible client

Every client needs the endpoint and Bearer header:

URL:    https://your-pulsyr-host.example/mcp
Header: Authorization: Bearer <PROJECT_TOKEN>

For Claude Code, the project settings screen provides a ready-to-copy command. Other MCP clients use the same URL and header in their own server configuration.

5. Verify project context

Restart the client if required, then ask it to call pulsyr_context. A valid response includes the project, current priorities, blockers, incidents, and active Threads. If authentication fails, create a new token from Project Settings instead of attempting to recover the old raw value.

Full MCP guideGet help on GitHub