SprintiQ provides an agile planning layer for AI-assisted coding workflows, specifically designed to work with tools like Claude Code. It handles sprint planning, user story generation, and task tracking while syncing changes bidirectionally with a local codebase. Built as a single-user, self-hosted application under the Apache 2.0 license, the project sits at https://github.com/SprintiQ-Incorporated/sprintiq with 37 GitHub stars. Written in TypeScript, it positions itself as the "product brain" for AI coding agents, managing what gets built, when, and why, rather than acting as a general project management tool.

The project addresses gaps in AI-driven development by generating user stories from ideas, planning sprints with capacity and velocity tracking, and linking them to actual code commits. It uses AI models like Claude Sonnet 4.6 and Opus via Anthropic, along with Voyage AI for embeddings, to create stories aware of agile anti-patterns through its TAWOS system. A CLI tool bridges the gap to local git repositories, allowing real-time mapping of commits to sprint tasks.

Core features

SprintiQ focuses on a narrow set of capabilities tailored to solo AI coding sessions:

  • Bidirectional sync via CLI: The sprintiq watch command tracks git activity in a local project and updates the corresponding sprint tasks.
  • AI user story generation: Produces multi-modal stories using persona-aware prompts and training on agile anti-patterns.
  • Sprint planning tools: Includes capacity management and velocity tracking for individual sprints.
  • Self-hosted data isolation: Runs on user-controlled Supabase instances with row-level security (RLS) for single-owner workspaces.
  • Storage for media: Supports public avatars bucket for profile photos and private images bucket for task uploads.

These features integrate directly with Claude Code, where the AI agent writes code while SprintiQ orchestrates the backlog.

Getting it running

Self-hosting requires a few external services but follows standard Node.js and Supabase patterns. Full details appear in SELF_HOSTING.md, but the quick start uses these steps.

Prerequisites

  • Node.js 18 or later.
  • A Supabase project (free tier works for personal use).
  • Anthropic API key for Claude Sonnet 4.6 and Opus.
  • Voyage AI API key for embeddings.

Web app setup

Clone the repo and configure:

git clone https://github.com/SprintiQ-Incorporated/sprintiq.git
cd sprintiq
cp env.example .env.local
# Edit .env.local with Supabase URL, Anthropic key, Voyage key, etc. (see SELF_HOSTING.md)
npm install
npx supabase db push
npm run dev

After supabase db push, create storage buckets in the Supabase dashboard:

Bucket name Public Purpose
avatars Yes User profile photos
images No Task and workspace images

The app uses Next.js App Router for pages and API routes, with Supabase handling auth, Postgres, and pgvector for vector search.

CLI setup for sync

The CLI lives in packages/cli and connects local code to the SprintiQ workspace.

cd packages/cli
npm install
npm run build
npm link

Authenticate once:

sprintiq login

Enter the API key from your SprintiQ workspace CLI settings. Then, from a git-initialized project root with an active sprint:

sprintiq watch

This command requires Node 18+, a .git folder, and an ongoing sprint. It maps commits to tasks in real time.

For development, additional commands include npm run typecheck for TypeScript checks, npm run lint for ESLint, npm run test for Vitest, and npm run test:e2e for Playwright.

Architecture overview

SprintiQ stacks Next.js on Supabase for the backend, with React components, contexts, and hooks in the frontend. Server utilities in lib/ manage AI providers. The CLI package enables the git bridge. Migrations sit in supabase/, types are shared across the monorepo, and tests cover unit, integration, and E2E cases. RLS ensures workspace data stays isolated to one user. CLAUDE.md offers a briefing for AI-agent integration.

Who this is for

Developers using Claude Code or similar AI agents for solo projects will find it useful. It fits workflows where you prompt an AI to generate code but need structure around stories, sprints, and progress tracking without team overhead. If you self-host other tools and manage your own Claude API keys, SprintiQ keeps data in your Supabase project. Single-user focus makes it ideal for indie hackers or solo consultants building apps iteratively. Those already tracking velocity manually or splitting stories via chat could automate that here.

It's less suited for teams, as RLS enforces single-owner isolation. General agile users without AI coding agents might skip it, since it lacks collaboration features found in broader tools.

Alternatives and comparisons

SprintiQ differs from traditional project management apps by tying directly to AI code generation via CLI sync. Tools like Jira or Linear offer sprint boards and velocity charts but require manual task updates and lack built-in AI story generation or git bridges. Trello or Notion provide flexible boards, yet they don't integrate with Claude Code or use embeddings for persona-aware planning.

Open-source options like Taiga or WeKan handle agile basics self-hosted, but without the AI layer or Claude-specific sync. For AI-focused planning, nothing matches its narrow integration; users might combine Obsidian plugins with git hooks for a lighter setup. SprintiQ carries more dependencies (Supabase, multiple APIs) than a pure frontend board, making it heavier for simple tracking.

Development and contributing

The monorepo supports standard npm scripts for linting, testing, and dev mode. External contributions follow CONTRIBUTING.md, with bug reports and security issues always accepted via SECURITY.md.

SprintiQ suits AI coding enthusiasts who self-host and want sprint orchestration without vendor lock-in. Source code and docs reside at https://github.com/SprintiQ-Incorporated/sprintiq.

View on GitHub →