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.
Related MCP server: agentbay-mcp
# Install the CLI (42KB, zero native dependencies)
npm install -g @sprintra/cli
# Authenticate
sprintra login
# Auto-configure your AI tool
sprintra connect# 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{
"mcpServers": {
"sprintra": {
"command": "npx",
"args": ["@sprintra/cli", "mcp"]
}
}
}Or run: sprintra connect --tool claude-code
Add to .cursor/mcp.json:
{
"mcpServers": {
"sprintra": {
"command": "npx",
"args": ["@sprintra/cli", "mcp"]
}
}
}Or run: sprintra connect --tool cursor
Add to ~/Library/Application Support/Claude/claude_desktop_config.json (macOS):
{
"mcpServers": {
"sprintra": {
"command": "npx",
"args": ["@sprintra/cli", "mcp"]
}
}
}Run: sprintra connect --tool windsurf
For remote MCP connections (Claude.ai, mobile):
URL: https://api.sprintra.io/api/mcp
Auth: Bearer <your-token>
Transport: HTTP + SSESprintra uses a consolidated tool pattern (inspired by GitHub's MCP server). Each tool accepts a method parameter to select the operation.
Tool | Methods | Description |
| list, create, update, resolve, get_context | Manage projects with tech stack, status, and full context snapshots |
| list, get_bundle, create, update, save_context, get_context | Epic-level work items with acceptance criteria and AI guidance |
| list, create, update, batch_update | Tasks under features — story, task, bug, chore types |
| list, get_current, create, update, assign | Time-boxed iterations with progress tracking |
| list, create, update, generate_notes | Release milestones with auto-generated release notes |
Tool | Methods | Description |
| list, add, supersede, get_conflicts, resolve_conflict | Architecture Decision Records (ADRs) with AI conflict detection |
| list, get, create, update, get_versions, restore_version, add_link, get_graph | Knowledge base with versioning, templates, and cross-references |
| list, add | Quick capture — ideas, research findings, meeting notes |
| list, start, log_message, complete | Brainstorming sessions with AI personas |
Tool | Methods | Description |
| add, remove, update_status, verify | Acceptance criteria management and verification |
| sync, link, unlink, get_status, get_feature_stats | Git commit tracking with AI authorship detection |
| add, remove, get_graph | Feature dependency tracking with cycle detection |
| list, start, end, get_active, list_activity, list_traces, delta | Coding session tracking with context handoff |
| list, create, update, delete | Threaded comments on any entity |
| manage, set, get | Custom metadata fields |
Tool | Methods | Description |
| get_next_work, report_progress, get_guidance, generate_briefing | AI-powered recommendations, progress tracking, and context recovery |
| list, get_story_prs | GitHub PR tracking and auto-linking |
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?"Install workflow skills for guided project management:
sprintra skills install --allSkill | Trigger | Description |
Capture Feature |
| Break down an idea into features, stories, and criteria |
Record Decision |
| Create an Architecture Decision Record |
Standup Report |
| Auto-generate standup from recent work |
Wrap Session |
| Save context for seamless session resume |
Sprint Review |
| Review sprint progress with metrics |
Brainstorm |
| AI-powered brainstorming with personas |
Implement Feature |
| Start implementation with full context loaded |
Sprint Planning |
| Plan sprints with AI recommendations |
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
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 |
Website: sprintra.io
Documentation: sprintra.io/docs
MCP Tools Reference: sprintra.io/docs/mcp-tools
Skills Guide: sprintra.io/docs/skills
CLI Reference: sprintra.io/docs/cli
npm: @sprintra/cli
Blog: sprintra.io/blog
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.
This server cannot be deployed
AI-native project management + agent memory: tasks, sprints, risk, burnout, knowledge search.
Work management where AI agents are first-class members: tasks, projects, memory over hosted MCP
Your versioned memory across every AI tool — context maps, personal memory, and tasks over MCP.
Project management for AI agents: tasks, docs, decisions and time in one shared team context.
