OpenRouter is the highway. Promptibus is the GPS.
OpenRouter is a brilliant unified API — one endpoint, 300+ models, transparent pricing per token. It solves the 'I need to call 5 different LLM APIs' problem. Promptibus MCP solves a different problem: WHICH of those 300 models should your agent pick for THIS specific task, and how should the prompt be formatted for it? We use OpenRouter pricing data in our routing logic. The two stack.
Use OpenRouter when you've already decided which model to call. Use Promptibus when you need to decide which model to call.
| Feature | Promptibus | OpenRouter |
|---|---|---|
| Primary function | Routing intelligence (which model + how) | Routing API (unified endpoint) |
| Output | Tool calls with optimized prompts | Forwarded LLM responses |
| Model decision | Agent calls `recommend_model({task})` | Agent picks model slug + sends prompt |
| Prompt optimization | Yes (`optimize_prompt` per model) | No — you send the prompt as-is |
| Pre-flight linting | Yes (`lint_prompt`) | No — you find out at generation time |
| Image / video / audio models | Yes (Midjourney, Flux, Sora, Suno, etc.) | Mostly text models |
| Free tier | 100 MCP calls/day | Free credits + per-call fees |
These are complementary. Use Promptibus `recommend_model` + `optimize_prompt` to figure out what to send, then send it via OpenRouter for the actual call. Promptibus's pricing data is curated independently and includes image/video/audio models OpenRouter doesn't.
Free tier: 100 calls/day. No credit card.