Sprintra MCP Server

AI-native project management with persistent memory for coding agents

Your AI forgets. Sprintra remembers.

What is Sprintra?

Sprintra gives your AI coding agent persistent memory — features, architecture decisions, sprint progress, and project context that survive session resets, token limits, and IDE switches.

The problem: Every AI coding session starts from scratch. Your agent doesn't know what you decided last week, what sprint you're in, or what your teammate built yesterday.

The solution: Sprintra's MCP server connects to Claude Code, Cursor, Windsurf, and any MCP-compatible tool. Your agent reads and writes project context automatically — no copy-paste, no re-explaining.

Quick Start

Option 1: Sprintra Cloud (Recommended)

# Install the CLI (42KB, zero native dependencies)
npm install -g @sprintra/cli

# Authenticate
sprintra login

# Auto-configure your AI tool
sprintra connect

Option 2: Self-Hosted

# Scaffold a complete local project with server + dashboard
npx create-sprintra my-project

# Start the server
cd my-project
npm start

# Server + dashboard at http://127.0.0.1:4000

MCP Configuration

Claude Code

{
  "mcpServers": {
    "sprintra": {
      "command": "npx",
      "args": ["@sprintra/cli", "mcp"]
    }
  }
}

Or run: sprintra connect --tool claude-code

Cursor

Add to .cursor/mcp.json:

{
  "mcpServers": {
    "sprintra": {
      "command": "npx",
      "args": ["@sprintra/cli", "mcp"]
    }
  }
}

Or run: sprintra connect --tool cursor

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):

{
  "mcpServers": {
    "sprintra": {
      "command": "npx",
      "args": ["@sprintra/cli", "mcp"]
    }
  }
}

Windsurf

Run: sprintra connect --tool windsurf

HTTP Transport (Remote)

For remote MCP connections (Claude.ai, mobile):

URL: https://api.sprintra.io/api/mcp
Auth: Bearer <your-token>
Transport: HTTP + SSE

17 MCP Tools

Sprintra uses a consolidated tool pattern (inspired by GitHub's MCP server). Each tool accepts a method parameter to select the operation.

Project Management

Tool	Methods	Description
`projects`	list, create, update, resolve, get_context	Manage projects with tech stack, status, and full context snapshots
`features`	list, get_bundle, create, update, save_context, get_context	Epic-level work items with acceptance criteria and AI guidance
`stories`	list, create, update, batch_update	Tasks under features — story, task, bug, chore types
`sprints`	list, get_current, create, update, assign	Time-boxed iterations with progress tracking
`releases`	list, create, update, generate_notes	Release milestones with auto-generated release notes

Knowledge & Decisions

Tool	Methods	Description
`decisions`	list, add, supersede, get_conflicts, resolve_conflict	Architecture Decision Records (ADRs) with AI conflict detection
`documents`	list, get, create, update, get_versions, restore_version, add_link, get_graph	Knowledge base with versioning, templates, and cross-references
`notes`	list, add	Quick capture — ideas, research findings, meeting notes
`sessions`	list, start, log_message, complete	Brainstorming sessions with AI personas

Tracking & Git

Tool	Methods	Description
`criteria`	add, remove, update_status, verify	Acceptance criteria management and verification
`git`	sync, link, unlink, get_status, get_feature_stats	Git commit tracking with AI authorship detection
`dependencies`	add, remove, get_graph	Feature dependency tracking with cycle detection
`work_sessions`	list, start, end, get_active, list_activity, list_traces, delta	Coding session tracking with context handoff
`comments`	list, create, update, delete	Threaded comments on any entity
`custom_fields`	manage, set, get	Custom metadata fields

AI Intelligence

Tool	Methods	Description
`ai`	get_next_work, report_progress, get_guidance, generate_briefing	AI-powered recommendations, progress tracking, and context recovery
`pull_requests`	list, get_story_prs	GitHub PR tracking and auto-linking

Example Prompts

Once connected, try these in Claude Code or Cursor:

"Create a Sprintra project for this repo"

"Capture a feature: user authentication with OAuth and email/password"

"What should I work on next?"

"Record a decision: we chose PostgreSQL over MongoDB for relational order data"

"Give me a standup report"

"What changed since my last session?"

8 AI Skills (Slash Commands)

Install workflow skills for guided project management:

sprintra skills install --all

Skill	Trigger	Description
Capture Feature	`/capture [idea]`	Break down an idea into features, stories, and criteria
Record Decision	`/decide [topic]`	Create an Architecture Decision Record
Standup Report	`/standup`	Auto-generate standup from recent work
Wrap Session	`/wrap`	Save context for seamless session resume
Sprint Review	`/sprint-review`	Review sprint progress with metrics
Brainstorm	`/brainstorm [topic]`	AI-powered brainstorming with personas
Implement Feature	`/implement [id]`	Start implementation with full context loaded
Sprint Planning	`/sprint-plan`	Plan sprints with AI recommendations

Dashboard

Sprintra includes a full web dashboard with 20+ views:

Kanban Board — drag-and-drop story management
Sprint Tracker — progress bars, velocity charts, burndown
Roadmap — phase timeline with feature grouping
Knowledge Base — Confluence-style docs with versioning
Decisions — ADR log with conflict detection
Reports — burndown, velocity, DORA metrics
Activity Feed — real-time project activity

Cloud: app.sprintra.io Self-hosted: http://127.0.0.1:4000 after npm start

Pricing

Plan	Price	Includes
Solo Pilot	Free forever	2 projects, unlimited stories, all MCP tools, all skills
Team	$5/seat/month	10 projects, AI features, team collaboration, priority support
Enterprise	Custom	On-premise, dedicated infra, custom SLA

License

The Sprintra MCP server documentation in this repository is MIT licensed. The Sprintra server, dashboard, and CLI are proprietary software — see sprintra.io/pricing for plans.

Sprintra