SprintiQ: Open-Source Sprint Planning for Claude Code

SprintiQ is a self-hosted, open-source (Apache 2.0) sprint planning tool built specifically for developers using Claude Code. It sits as an orchestration layer above Claude Code: while Claude writes code, SprintiQ manages what gets built, when, and why. It is not a generic project management tool — it's an operating system for Claude Code workflows.
Key Features
- Bidirectional sync with Claude Code via the SprintiQ CLI (
sprintiq watch) - AI-powered user story generation trained on agile anti-patterns (TAWOS)
- Sprint planning, capacity management, and velocity tracking
- Persona-aware story generation — adapts stories based on defined user personas
- Single-user, self-hosted — your data, your infrastructure, your Claude API key
- Multi-modal story generation (text + image support via Supabase storage)
Quick Start (Self-Hosted)
Prerequisites:
- Node.js 18 or later
- A Supabase project (free tier works for personal use)
- An Anthropic API key (Claude Sonnet 4.6 + Opus)
- A Voyage AI API key (for embeddings)
Setup:
git clone https://github.com/SprintiQ-Incorporated/sprintiq.git
cd sprintiq
cp env.example .env.local
# Fill in the required env vars — see SELF_HOSTING.md
npm install
npx supabase db push
npm run dev
After supabase db push, create two storage buckets in your Supabase dashboard:
avatars(public) — for user profile photosimages(private) — for task and workspace image uploads
CLI: sprintiq watch
The CLI connects your local codebase to your SprintiQ workspace, tracking git activity and syncing it to your active sprint in real time.
Install and authenticate:
cd packages/cli
npm install
npm run build
npm link
From your project directory:
sprintiq login
Start watching:
sprintiq watch
Requirements:
- Must run from a git-initialized project directory
- Requires an active sprint in your SprintiQ workspace
- Node 18+
Architecture
Built on Next.js App Router, Supabase (auth, Postgres, pgvector), Claude Sonnet 4.6 for generation, and Voyage AI for embeddings. The CLI creates a live bridge between Claude Code sessions and your sprint board. Row-Level Security (RLS) enforces single-owner workspace isolation at the database layer.
The repo includes a CLAUDE.md file optimized for AI agent context — useful if you want Claude Code to understand the project's structure.
Who It's For
Developers who use Claude Code for development and want a lightweight, self-hosted planning layer that syncs git activity with sprint tasks — without leaving the terminal or sending data to a third-party SaaS.
📖 Read the full source: HN AI Agents
👀 See Also

Lore: MCP Server That Shares AI Agent Session History Across Tools
Lore is an MCP server that indexes AI agent sessions into a local SQLite store, letting any agent—regardless of tool—access another's session history. Fresh client sessions, no shared context, yet agents can pull up past conversations on demand.

Nelson v2.2.3 Released: Multi-Agent Coordination for Claude Code, Plus a Discrete-Event Simulation Benchmark
Nelson v2.2.3 ships a multi-agent coordination skill for Claude Code using a naval metaphor. A 13-configuration benchmark shows opus-4-7 with thinking dominates; skill choice is a smaller delta.

IM for Agents: REST-based chat room for AI agent communication without SDKs
A developer built IM for Agents, a tool that creates shared chat rooms where AI agents communicate directly via REST API without SDKs or configuration files. Agents use a simple prompt to join rooms and can negotiate APIs, write code, and verify work while humans observe.

Get Shit Done: Meta-Prompting System for AI Coding Agents
Get Shit Done is a meta-prompting, context engineering, and spec-driven development system that works with Claude Code, OpenCode, Gemini CLI, Codex, Copilot, and Antigravity. It addresses context rot by providing structured prompts and verification workflows.