Log inSubmit server
Saga MCP logo

Saga MCP

by spranab

SQLite-backed MCP project tracker with dependencies, comments, templates, dashboards, and tools for managing work across sessions.

InstallRepositoryWebsite
Full README Compare alternatives Embed widget Claim this listing

Install Saga MCP

Pick your client, then add it in one click or paste the config.

STDIONode.js
~/Library/Application Support/Claude/claude_desktop_config.json
claude_desktop_config.json
{
  "mcpServers": {
    "spranab-saga-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "-p",
        "saga-mcp",
        "saga-web"
      ],
      "env": {
        "DB_PATH": "YOUR_VALUE_HERE",
        "SAGA_PROJECT": "YOUR_VALUE_HERE",
        "SAGA_DESCRIPTION_LOCK": "YOUR_VALUE_HERE",
        "SAGA_TOOLS": "YOUR_VALUE_HERE"
      }
    }
  }
}

Settings → Developer → Edit Config, merge this into mcpServers, then fully quit and reopen Claude Desktop.

Values are written into the snippet above so you can paste it as is. They never leave your browser.

We couldn’t confirm it starts correctly

We ran the install command but it didn't respond within our test window — this can mean a slow first-time install rather than a real problem.

npx -y -p saga-mcp saga-web

No response to initialize.

This is an experimental automated check and can have false negatives — missing environment variables, a slow cold install, etc. It doesn’t necessarily mean something’s wrong. Last checked 24d ago.

Overview

spranab/saga-mcp MCP server gives AI agents a local project tracker backed by a SQLite database. It models projects, epics, tasks, subtasks, dependencies, threaded comments, reusable templates, and activity history through MCP tools. Dependency changes can automatically block or unblock work, while dashboard and next-task tools provide status and prioritization context. Reach for it when an agent needs durable planning across sessions without an account, hosted service, or network connection.

Use cases

  • Track multi-session coding projects
  • Plan epics and dependent implementation tasks
  • Resume work with a reasoned next-task recommendation
  • Record threaded decisions and progress notes
  • Reuse task sets with variable-based templates

Key features

  • SQLite-backed local persistence
  • Project, epic, task, and subtask hierarchy
  • Automatic dependency blocking and unblocking
  • Natural-language dashboard and next-task recommendations
  • Threaded comments and activity logging
  • Reusable templates with variable substitution

Tools41

What Saga MCP lets your agent do. Expand a tool for its parameters.

Self-reported~568 tokens of context

Security & health

What it can reach and how reliably it runs — from our own checks, not the maintainer’s claims.

ActiveLast checked 1d ago

Recent health check succeeded.

18/19 checks healthy over the last 79d

Supply-chain signal

No high-severity advisories surfaced by our automated scan.

Critical 0High 0Medium 0Low 0

Scanned 7d ago via OSV.dev · saga-mcp (npm)

Adoption & maintenance

Factual signals from GitHub, npm, and our automated checks — not a rating.

GitHub stars40Stargazers on the source repository.
npm downloads1.9kPackage downloads in the last 30 days.
Last commit21d agoMost recent push to the default branch.
Install checkInconclusiveDidn't respond in our test window — often a slow first install.
Tools exposed41Callable tools this server registers over MCP.
Directory activity4 viewsConfig copies, upvotes, and views on AllMCPs.

How Saga MCP works

What spranab/saga-mcp MCP server does

spranab/saga-mcp MCP server stores project planning data in a SQLite file that belongs to the user. Its hierarchy runs from projects to epics, tasks, and subtasks, allowing an agent to preserve work plans instead of relying on temporary conversation context.

The tracker supports task dependencies, threaded comments, reusable templates, manual ordering, soft deletion, restoration, archiving, and activity logging. Task updates can record changes and recalculate dependency status. A task description can also be locked to prevent an agent from replacing the original specification with progress notes.

How it works

The MCP process reads and writes the configured SQLite database. tracker_init can create the initial project, while project, epic, task, subtask, comment, and template tools manage the stored records. The database file and schema are created on first use, so no external database service is required.

Dependencies are evaluated when relevant task state changes. A dependent task becomes blocked while prerequisite work is unfinished and can become available again when the prerequisite is completed. Reopening a prerequisite can block its dependents again, and circular task or subtask dependencies are rejected.

tracker_next selects a recommended next task and explains the choice, including blocked or overdue work. tracker_dashboard returns a broader project view with statistics, epics, blocked and overdue tasks, recent activity, notes, and a natural-language summary. Task lists follow a manually assigned epic order when one exists; otherwise, priority sorting applies unless another sort is requested.

Setup and configuration

Add the server to an MCP client using npx -y saga-mcp and set DB_PATH to an absolute path for the SQLite file. The same configuration format is documented for Claude Code and Claude Desktop, and the package can also be installed globally with npm before using saga-mcp as the command.

DB_PATH is required. SAGA_PROJECT optionally scopes tools to a project by ID or name, which is useful when several repositories share one database. SAGA_DESCRIPTION_LOCK can enable automatic description locking when work starts. SAGA_TOOLS accepts full or core; the latter exposes a smaller tool set for ordinary tracking sessions.

The README reports testing with Node 20, 22, and 24 on Linux, macOS, and Windows. No API key, account, or network service is required.

Tools and capabilities

The listed tools cover:

  • Creating, listing, updating, archiving, and restoring projects, epics, tasks, and subtasks.
  • Adding task dependencies, reordering tasks or subtasks, and applying batch updates or deletions.
  • Locking task descriptions and retrieving task details with subtasks, notes, comments, and dependencies.
  • Adding, listing, deleting, and restoring threaded task comments.
  • Creating, listing, and editing templates with {variable} placeholders.
  • Checking the next recommended task and viewing a project dashboard.

Limitations and notes

spranab/saga-mcp is local-file based, so the database path must be available to the MCP process. It does not provide an external hosted service or network synchronization. Archived epics and soft-deleted records are hidden from normal listings, although supported restore operations can recover removed tasks and comments.

Read the full README View source on GitHub

Reviews

No reviews yet — be the first to share how it worked for you.

Frequently Asked Questions about Saga MCP

spranab/saga-mcp is an MCP server that connects AI agents to a local SQLite project tracker. Its main tools create and manage projects, epics, tasks, subtasks, dependencies, comments, and templates, while tracker_next recommends what to do next and tracker_dashboard summarizes project status without requiring an account or external service.

Install it by adding an MCP configuration that runs `npx -y saga-mcp` and sets the required `DB_PATH` environment variable. You can also install it globally with npm and use `saga-mcp` as the command. Restart the MCP client after changing its configuration.

The required variable is `DB_PATH`, which identifies the SQLite database file. The file and schema are created on first use. Optional variables include `SAGA_PROJECT`, `SAGA_DESCRIPTION_LOCK`, and `SAGA_TOOLS`.

Yes. The README provides a Claude Desktop configuration using the server command and `DB_PATH`. It also documents the same style of configuration for Claude Code and says the server can be used with other MCP clients.

It automatically blocks tasks whose dependencies are unfinished and unblocks them when prerequisites become complete. Reopening a completed prerequisite can block dependents again, while circular dependencies are refused. Subtasks support dependency relationships as well.

Related MCP servers

All in Product Management All alternatives
Add a badge to your READMEShow it’s listed — live SVG, Markdown or HTML.

AllMCPs Directory Badge

Full Badge Customizer

Showcase your server listing on GitHub or your project documentation. Embed this dynamic SVG badge to highlight official listing status and live engagement.

Badge Style:
Live Dynamic SVG PreviewSaga MCP AllMCPs Directory Badge
Markdown (GitHub README)
[![AllMCPs](https://allmcps.com/api/badge/spranab-saga-mcp?style=directory)](https://allmcps.com/mcp/spranab-saga-mcp)
HTML Embed
<a href="https://allmcps.com/mcp/spranab-saga-mcp"><img src="https://allmcps.com/api/badge/spranab-saga-mcp?style=directory" alt="Saga MCP on AllMCPs" /></a>