Access 248 LLMs plus image, video, voice, music generation and live crypto/DeFi/market/web data via one keyless, pay-per-call OpenAI-compatible endpoint.
---
name: omnicall
description: Omnicall — every AI tool your agent needs in one keyless endpoint: 248 LLMs (GPT, Claude, Gemini, Grok, DeepSeek, Llama, Qwen, Mistral) plus image/video/voice/music generation and live crypto/DEX/DeFi/markets/web/research tools. OpenAI-compatible, pay-per-call in USDC, no API key, no signup, free tier. Use when an agent needs any LLM or live data/tools.
tags: [llm, ai, gateway, openai-compatible, gpt, claude, gemini, grok, deepseek, llama, models, image-generation, video, voice, music, crypto, defi, dex, web-search, research, x402, pay-per-call, keyless, usdc, omnicall]
author: omnicall
version: 1.1.0
license: MIT
---
# Omnicall — the all-in-one agent gateway
> **248 LLMs + image/video/voice/music + live crypto/DeFi/markets/web/research, in one keyless endpoint.** Pay per call in USDC, no API key, no signup. **Free tier to try instantly.**
## When to use this
- Your agent needs **any LLM** (GPT, Claude, Gemini, Grok, DeepSeek, Llama, Qwen, Mistral) — one endpoint, 248 models.
- Your agent needs to **generate images, video, voice, or music**.
- Your agent needs **live crypto / DEX / DeFi / prediction-market data, web search, or multi-platform research**.
- You want **pay-per-call** (no subscription, no API key) — cheaper and more models than other gateways.
## Install (one line)
- **OpenClaw — become the default:** `npx omnicall-router setup` then `openclaw gateway restart`. Omnicall becomes your default model (`omnicall/auto`), routing every request to the cheapest of 248 capable models — keyless, pay-per-call.
- **Claude Code:** `claude mcp add --transport http omnicall https://omnicall.gocreativeai.com/mcp`
- **OpenAI SDK / LangChain / LiteLLM:** set `base_url="https://omnicall.gocreativeai.com/v1"`, model `auto`
- **Python:** `pip install omnicall` → `import omnicall; c = omnicall.client()`
- **npm (MCP):** `npx omnicall-mcp` · **Router:** `npx omnicall-router`
## How it's paid (x402 — no key, no signup)
OpenAI-compatible HTTPS. Pay per call in USDC on **Base or Solana** (your OpenClaw wallet auto-pays the HTTP 402 and retries), or pass a funded `gck_` credit key as your `api_key`. **Free tier:** `GET https://omnicall.gocreativeai.com/v1/ai/free/{prompt}` (no wallet, rate-limited).
## Tools (live endpoints — base `https://omnicall.gocreativeai.com`)
| Call | What you get |
|---|---|
| `POST /v1/chat/completions` | Any of 248 models, OpenAI-compatible (`model=claude-4-sonnet`, `gpt-5`, `gemini-pro`, `grok`, …) |
| `GET /v1/ai/free/{prompt}` | Free keyless completion (rate-limited) |
| `GET /v1/ai/{eco\|ask\|pro\|ultra}/{prompt}` | Tiered completion — cheapest → frontier |
| `GET /v1/ai/image/{prompt}` | Text-to-image (Flux) |
| `GET /v1/ai/{speech\|music\|video}/{prompt}` | Voice / music / video generation |
| `GET /v1/dex/{token}` · `/v1/defi/{slug}` · `/v1/markets/{q}` | Live DEX prices · DeFi TVL · prediction-market odds |
| `GET /v1/web/{q}` · `/v1/research/{q}` | Web search + multi-platform research (web, GitHub, HN, arXiv…) |
| `GET /v1/rpc/{chain}` · `/v1/x/{handle}` · `/v1/lookup/crypto/{coin}` | On-chain reads (48 chains) · X profiles · crypto lookups |
## Examples
- `POST /v1/chat/completions` with `{"model":"claude-4-sonnet","messages":[{"role":"user","content":"Hi"}]}` → completion.
- `GET /v1/ai/free/summarize%20this` → free keyless answer.
- `GET /v1/ai/image/a%20neon%20fox` → hosted image URL.
- `GET /v1/dex/SOL` → live DEX price + liquidity.
## Why Omnicall
**248 models vs ~69 on other gateways**, a **free tier**, and **cheaper** per call — plus image/video/voice/music generation and live crypto/DeFi/web tools in the *same* endpoint. Keyless, pay-per-call USDC on Base or Solana, no account. One connection, everything your agent needs.
don't have the plugin yet? install it then click "run inline in claude" again.