sally docs

Documentation for humans, scripts, and agents.

Sally is an API-first project management system with a low-noise web UI, a real HTTP API, and MCP support for agent workflows. This docs hub mirrors the current product shape and points to the implementation-backed reference docs in the repo.

Install
Ubuntu/Debian installer, managed-simple, existing-infra, release notes.
Operate
Workspaces, projects, tasks, comments, notifications, timesheets, recovery.
Integrate
API auth, workspace selection, hosted MCP, stdio MCP, examples.
Start here
  • Install: npx --yes create-sally@latest
  • API reference: docs/api.md
  • MCP guide: docs/mcp.md
  • Product/workflow guide: docs/product-guide.md
  • Tutorials + examples: docs/tutorials.md
Recommended order for new operators:
  1. README
  2. Product guide
  3. Install guide
  4. MCP guide
  5. Recovery notes
Current app state
  • Workspaces, memberships, invites, and role-based access
  • Projects with statuses, activity, and project memberships
  • Tasks with labels, todos, comments, due dates, and inline images
  • Clients, notifications, and timesheets
  • Hosted MCP at /mcp plus local stdio MCP via sally-mcp
The implementation-backed source of truth is the API code plus the repo docs, especially docs/api.md.
Hosted MCP quick example

Hosted MCP is a first-class Sally feature. Typical flow: initialize, send notifications/initialized, list tools, then call tools with the returned session id.

curl -X POST https://your-sally-domain.com/mcp   -H 'Authorization: Bearer sallymcp_...'   -H 'Content-Type: application/json'   -H 'Accept: application/json, text/event-stream'   --data '{
    "jsonrpc":"2.0",
    "id":1,
    "method":"initialize",
    "params":{
      "protocolVersion":"2025-03-26",
      "capabilities":{},
      "clientInfo":{"name":"manual-test","version":"1.0.0"}
    }
  }'
Useful repo docs
  • docs/index.md
  • docs/product-guide.md
  • docs/api.md
  • docs/mcp.md
  • docs/tutorials.md
  • docs/ubuntu-debian-install.md
  • docs/recovery.md
Documentation intent
These docs are written to be readable by humans and dependable for LLMs. That means:
  • explicit terminology
  • practical examples
  • clear auth/workspace selection guidance
  • reference docs that stay close to the implementation