GeiserX/LynxPrompt

Self-hostable config hub for sharing AI IDE settings across teams.

What it does

• Config generation wizard — step-by-step builder (web + CLI) for CLAUDE.md, AGENTS.md, .cursor/rules, and 30+ other AI assistant formats
• Blueprint marketplace — private or federated library for sharing reusable team configurations
• Multi-tool export — write once, export to Cursor, Claude Code, GitHub Copilot, Windsurf, Zed, Cline, Junie, and others
• REST API + CLI — programmatic access for CI/CD integration and automation
• Authentication menu — OAuth (GitHub/Google), magic-link email, WebAuthn passkeys, SAML/OIDC/LDAP
• Self-hosted or SaaS — Docker Compose stack or managed instance at lynxprompt.com

How it works

Next.js web app + Node.js backend + PostgreSQL. Auto-detects tech stacks from GitHub URLs. Supports template variables and monorepo hierarchies. Optional Claude AI-assisted editing when ENABLE_AI=true + Anthropic API key provided. Compose + Helm charts for deployment; all features toggled via environment variables.

Standing & signals

43 stars and last-pushed 2026-08-10. Distributed across npm, Chocolatey, Docker Hub, AUR, Snap, and Helm Chart — suggests author has infrastructure maturity even with modest adoption. No release dates in the README; single-author project with a public SaaS instance.

Q2 — Apply where

Not now. No immediate fit to active projects. Configuration file generation is orthogonal to ClientCo's contract-first intelligence engine, Financial Analysis, or Research Console's session + routing layer.

Q3 — Fit for current projects + sketch

If the Research Console expands to team/org-scale governance (currently per-operator, in-process), LynxPrompt could become the config-sharing layer. Today, Research Console embeds constitutions in YAML; a future org mode might delegate to a blueprint marketplace. Sketch: (1) export constitution as LynxPrompt blueprint, (2) team members fork + customize, (3) sync back to console on session start. But this is speculative; the console's current design is anti-centralized (per-project, held calls resolved mid-turn by operator).

Q4 — Future-watch verdict

IGNORE. Minimal adoption (43⭐), solves a team/org problem the owner does not yet have, and the governance pattern it enables diverges from Research Console's operator-centric architecture.