5 agent-payable APIs. Every listing carries a tested call recipe - endpoint, auth shape, pricing, and known gotchas - on its detail page. erb-llm.com ↗
OpenAI-compatible text embeddings from a private, self-hosted local encoder (768 dims). Pay per call in USDC on Base via x402 - no API key or account. Send {"input": string or array of up to 64 texts}; receive standard OpenAI embeddings JSON with usage. Built for RAG, semantic search, dedupe and clustering in agent pipelines. ~2,000-token limit per text. Same privacy-first operator as our Private LLM Inference chat tiers.
OpenAI-compatible vision (image understanding) from a private, self-hosted local vision-language model. Pay per call in USDC on Base via x402 - no API key or account. Send a public image URL plus a prompt (or OpenAI image_url messages); receive a standard chat.completion JSON answering about the image. Up to 4 images per call (more are clamped). Same privacy-first operator as our Private LLM Inference chat tiers.
OpenAI-compatible chat completions served by a self-hosted 27B open-weight LLM. Pay per call in USDC on Base via x402 - no API key or account. Use it for text generation, summarization, classification and agent tool calls that need cheap, private inference. Send a messages array (plus optional model, max_tokens, temperature); receive standard OpenAI chat.completion JSON. A free trial (POST /v1/trial) is open during promo windows.
Extended tier of the x402 LLM gateway: same self-hosted 27B open-weight LLM with the highest output ceiling (32,768 tokens) for long-form writing, reports and full document drafts. Pay per call in USDC on Base via x402 - no API key or account. OpenAI-compatible chat completions. Pick the cheaper quick/long tiers for shorter work.
Long-form tier of the x402 LLM gateway: same self-hosted 27B open-weight LLM with an 8,192-token output ceiling for summarization, drafting and multi-paragraph answers. Pay per call in USDC on Base via x402 - no API key or account. OpenAI-compatible: send a messages array, receive standard chat.completion JSON. Use the cheaper /v1/chat/completions tier (4,096 tokens) for short answers and tool calls.