Common questions about the tools, the stack, privacy, and how everything is wired up.
Yes. Every AI tool on the site is free — no sign-up, no subscription, no API keys required. You can open the chat, bug fixer, image generator, or any other tool and use it immediately.
Vibecoding is the practice of building software by describing what you want in natural language and letting an AI assistant handle the code. You focus on intent and iteration; the model fills in the implementation. RnR Vibe is designed around this workflow.
No. There is no account system. Rate limiting is applied per-IP to prevent abuse, but nothing gates the tools behind a login or email.
Requests route through a local Ollama-hosted model (Gemma 3 / Gemma 3n class) on the RnR Vibe host machine. When the local machine is offline, requests transparently fall back to a free OpenRouter model so the tools stay available.
Image tools (Image Generator, Image Studio, Logo Generator) call a local Stable Diffusion WebUI (AUTOMATIC1111) running on the host machine. When the GPU service is offline, the tool shows a banner and lets you retry.
No. Your prompts are not persisted to a database and are not used to train any model. Request metadata (timestamp, IP, prompt hash) is logged for rate limiting and abuse prevention, and is rotated out.
The code is open on GitHub and the stack is designed to be self-hostable. You need Node.js, Ollama, and optionally AUTOMATIC1111 for image tools and the audio server for music/SFX. See the Guides section for setup walk-throughs.
A background health probe polls Ollama every 30 seconds. If it stops responding, the chat API routes new requests to OpenRouter instead. When Ollama comes back, the next successful probe flips the switch back. The swap is invisible to the user beyond a model-name difference in advanced debug output.
Image tools only run when the host machine is online and Stable Diffusion is started. SD is VRAM-heavy so it is started on demand rather than left running. If you see the banner, it usually clears within a minute or two — the retry button re-probes the service.
Each tool has its own per-IP rate limit window in memory. The Deep Research tool is stricter (roughly 10 requests per hour per IP) because each request fans out to ~10 external HTTP calls. Most other tools allow several requests per minute. You will see a 429 response if you exceed the limit.
There is no public API or SDK. The tools are intended as a browser-first vibecoding sandbox. If you want programmatic access, self-host the repo — every tool is a thin wrapper over Ollama and Stable Diffusion, both of which expose HTTP APIs.
Those are IDE-class agents designed for multi-file edits on a full codebase. RnR Vibe is a lightweight, browser-based sandbox with narrow, single-purpose AI tools (regex generator, SQL generator, image generator, bug fixer, etc.) that you reach in one click without installing anything. See the /compare pages for head-to-head breakdowns.
No. The site is responsive and works well on mobile browsers, but there is no dedicated iOS or Android app.
RnR Vibe is run by a small team and issues/feature requests go through GitHub. The About page links the contact path. Critical bugs (tool returning wrong output, broken pipeline) are prioritized over cosmetic requests.