HivePaaS
An open-source, self-hosted PaaS on Docker Swarm. Deploy apps, databases and functions from one VPS to a whole cluster, with SSL, backups and autoscaling built in.
Turn an idea, a research paper or a photo of hardware into a working prototype.
Translation pending
Developer toolsTypeScript · PythonGitHub @openideaorgOnlineOwner unconfirmed
Have you tried it? One tap is enough. No sign-in needed.
No updates detected yet. GitHub releases will appear here.
Owners of public repositories under a personal GitHub account can sign in to verify ownership and manage the introduction themselves. No administrator approval is needed.
Sign in, generate a verification file, and commit it to this repository’s default branch within 30 minutes. Files on other branches or forks do not count. Then check the file below to manage the project without administrator approval. Keep the file to retain your access.
If needed, you will go to sign-in. After signing in, return here and press your chosen verification button again.
Run /nomorevibe verify in your project folder to verify ownership and manage the introduction yourself.
Verify in one minuteAsk about other verification options
Turn an idea, a research paper, or a hardware photo into a working prototype. Muse Spark + SAM 3.
Translation pending
Taken from the repository README
Loading the full README.
Read the full README ↗OpenIdea Lab turns an idea or research paper into a verified, citable prototype with a Muse-only agent loop (plan / act / observe), HITL sign-off, and Studio handoff.
Built during the Meta Global AI Developer Hackathon (Oct 26 – Nov 4, 2026), Startup track. Everything in this repository was written during the event. It calls OpenIdea's search API (openidea.world), which existed before the hackathon and is not part of this submission. See PARTNERS.md.
| SKU | Price | What you get |
|---|---|---|
| Free | $0 | Teaser Lab Agent runs (metered), live console |
| Lab Pro | ≤ $20/mo | More runs, durable replay fence, BibTeX export, HITL, Studio files handoff |
Primary UI: /agent — Muse chooses one tool per turn. Legacy /idea and /paper one-shot pipelines remain as demos. Photo / SAM 3 on Baseten is optional vision GPU, not LLM reasoning.
Lab does not invent API keys. The Muse-only path reads:
META_API_KEY=… # from Configure Claude Code / Cursor-configured Muse API keys
META_API_URL=https://api.meta.ai/v1
# Router (lib/muse-models.ts): Wonder/Spark primary builders; Contributor for light steps + cheap retries
# choose/plan/verdict → muse-wonder-1.0 | draft/fix → muse-spark-1.3
# rank/observe → muse-spark-1.2-contributor | enrich/vision → muse-spark-1.3-contributor | segment → sam-3.1
Wire path: .env.local → lib/env.ts keyFor('meta') → lib/muse-models.ts routeMuseModel(task) → lib/muse.ts / lib/muse-agent.ts → Meta Model API.
idea / paper
│
▼
Muse agent loop (one tool / turn)
search_open · expand_graph · set_plan · draft_code
run_sandbox (browser) · fix_code · judge_paper
cite_sources · submit_prototype · request_human · finish
│
▼
console (plan / act / observe) + usage (Meta) → export / Studio
Details in docs/ARCHITECTURE.md.
git clone https://github.com/openideaorg/openidea-lab
cd openidea-lab
pnpm install
cp .env.example .env.local # set META_API_KEY from Cursor/Claude Code Muse keys
pnpm dev
Open /agent. Optional: LAB_PLAN=pro and LAB_PRO_PRICE_USD=20 (hard-capped at 20).
See CONTRIBUTING.md. Security: SECURITY.md.
MIT.