LLM inference APIs for AI agents
LLM inference APIs sell a model call the way other listings sell a lookup: one request, one price, no monthly seat. This page keeps Intelligence, Tools and Data listings that say LLM, chat completion, language model or embedding, or whose name says inference or GPT. A summary of a news article is not included just because it summarises something. The agent finds the listing through the free catalogue, then pays the provider over x402 in USDC on Base, Solana or Algorand. No API key is involved. Settlement-verified means a paid route with an on-chain receipt, and the examples are claimed listings that have one. Unclaimed (bazaar) imports are labelled and sit below the examples. Listing a model endpoint is free and the provider keeps 100% of each settled call. API and endpoint counts are read from the live catalogue with this rule, so they are not a forecast.
29 APIs and 266 endpoints match this page. 29 are settlement-verified: a paid route with an on-chain receipt.
Settlement-verified examples
- Anchor-x402 · $0.001 to $1.77 · settlement-verified · On-chain receipt
16 x402-paid endpoints for AI agents: OFAC sanctions screening, wallet due-diligence, dual-chain hash anchoring, signed receipts, tx/calldata decode, ENS/SNS, prices + LLM tools. USDC, no keys. - AARN Web Content Extraction · $0.005 · settlement-verified · On-chain receipt
Extract clean, LLM-ready Markdown or plain text from public webpages for AI agents, RAG, research and summarization. Pay-per-call via x402 on Base. - Agent Extraction API: LLM token + cost calculator across models · $0.005 · settlement-verified · On-chain receipt
Token & cost calculator: estimate tokens for a text (or pass input_tokens/output_tokens) and price it across several models using provider list prices (LiteLLM price table, MIT). Params: models (comma list of model ids like gpt-4o-mini, claude-sonnet-4-5, gemini/gemini-2.5-flash or search words, max 15), text or input_tokens, output_tokens, calls. - Agent Extraction API: Page to RAG (web page to Markdown + chunks) · $0.0025 · settlement-verified · On-chain receipt
Any public web page to clean LLM-ready Markdown plus heading-aware RAG chunks in one call: Mozilla Readability main content, absolute links, title/description/language/canonical/dates/JSON-LD types, word and token estimates, sha256 content hash for change detection. Respects robots.txt. $0.001 USDC per call on Base; failed calls are not charged; 1 free call per caller. Docs: https://freetier-agent-api.pages.dev/products/page-rag - Agent Trust, Verification & Machine-Payable Data Utilities API · $0.02 to $2 · settlement-verified · On-chain receipt
Machine-payable verification and trust services for autonomous agents on x402 (USDC on Base). An independent quality-control layer for agents that buy from other agents: submit a third-party output and receive ONE Ed25519-signed, SHA-256-reproducible verdict covering schema, freshness, internal consistency, evidence, and reference-checkable accuracy. Also: conformance-audit any x402 endpoint against a versioned published standard; conformance-check any issuer's signed receipt or payment offer; compute deterministic issuer reputation from signed observation history; govern agent spend with atomic reserved leases, settle and receipts; attest execution readiness and containment; verify C2PA-style content provenance against a content hash; seal caller-supplied data snapshots with honest freshness tiers; store and query a signed evidence vault; prove identity control; threat-model tool permissions; grade model output pairs against a deterministic rubric; and extract invoices, SEC filings, documents and redacted PII. Every paid route answers HTTP 402 with a USDC-on-Base envelope; every signature is verifiable free at a companion route. Deterministic engines only: no LLM inference, no invented scores, and any check that cannot be assessed is named honestly instead of guessed. - AgentDiet Semantic Context Compressor & Token De-Bloater · $0.001 to $0.025 · settlement-verified · On-chain receipt
Strips 85-97% of HTML, DOM, and JSON token bloat into dense LLM-ready Markdown with SHA-256 ROI proof for AI agents. - AgentJSON Fix · $0.005 · settlement-verified · On-chain receipt
Fix broken LLM/agent JSON into valid JSON — optional schema check — $0.005 USDC per call. - Flat-rate LLM and EVM toolkit · $0.005 · settlement-verified · On-chain receipt
48 pay-per-call endpoints, no API key and no account: EVM chain reads, ERC20 safety checks, URL to markdown, email and domain checks over live DNS, hashing, randomness, time zones and LLM completions. Verified route: GET /v1/evm/block?chain=base at $0.005, settled in USDC on Base. That route also advertises Polygon and Arbitrum payment rails at the same price; those rails are not verified. The other endpoints on this host, including the $1 completions route, are listed but unpaid and unverified. - FreeTier Agent API · $0.0015 · settlement-verified · On-chain receipt
24 x402 pay-per-call utility & data routes for AI agents: public holidays, country/currency data, Base gas, LLM pricing/cost, MCP server & x402 service directories, free-tier API search, page markdown/meta/link/robots checks, email verify, JSON extraction. $0.001 USDC per call on Base (6 bulk dataset dumps $1). 1 free call per route. OpenAPI: https://freetier-agent-api.pages.dev/openapi.json - Gondola LLM Inference API · $0.01 to $1 · settlement-verified · On-chain receipt
Pay-per-request LLM inference over x402. 350+ Venice AI models incl. uncensored lines. OpenAI chat + images, Anthropic messages. Every 402 quotes the exact price; overpayment banks as credit. - IA Inference API · $0.001 · settlement-verified · On-chain receipt
OpenAI-compatible chat completions served by local GPU models in Quebec, Canada. Pay per call in USDC on Base. - JSON-Repair Deterministic Hallucinated JSON Healer · $0.001 to $0.025 · settlement-verified · On-chain receipt
Deterministic syntax healer and schema normalizer for malformed or truncated LLM JSON outputs.
Unclaimed (bazaar) imports
These rows were imported and the provider has not claimed them. They are not the examples above.
- Chat completions via Telnyx Inference (google/gemma-2b-it) · $0.002 · settlement-verified · Unclaimed (bazaar)
Chat completions via Telnyx Inference (google/gemma-2b-it). OpenAI-compatible. $0.002 USDC per call. - Fetch any public URL, render client-side JS, strip nav/ads, return clean… · $0.002 · settlement-verified · Unclaimed (bazaar)
Fetch any public URL, render client-side JS, strip nav/ads, return clean LLM-ready Markdown. Fetch any public URL, render client-side JS, strip nav/ads, return clean LLM-ready Markdown. - Read any live web page as clean, LLM-ready text · $0.001 · settlement-verified · Unclaimed (bazaar)
Read any live web page as clean, LLM-ready text. Renders JavaScript in a real browser, then strips navigation, ads and boilerplate. Use it when a plain HTTP GET returns empty or garbled HTML: single-page apps, dynamic feeds, or content that only appears after scripts run. Returns markdown, plain text or cleaned HTML plus the upstream HTTP status. Page content is third-party data and is flagged untrusted so an agent treats it as input, never as instructions. - URL to Markdown API and converter for AI agents — convert any public HTTPS URL… · $0.001 · settlement-verified · Unclaimed (bazaar)
URL to Markdown API and converter for AI agents — convert any public HTTPS URL to clean Markdown, HTML-to-text, and LLM-ready web content with no separate scraper account or API key. URL to Markdown API and converter for AI agents — convert any public HTTPS URL to clean Markdown, HTML-to-text, and LLM-ready web content with no separate scraper account or API key.
Related pages
Questions
Which listings count as LLM inference?
Intelligence, Tools and Data listings that mention LLM, chat completion, language model or embedding, or whose name says inference or GPT. Tags llm, inference and embedding match as well.
How does an agent pay for a model call?
The same way it pays for any other listing. The provider returns HTTP 402, and the agent pays in USDC on Base, Solana or Algorand over x402. No API key and no subscription.
Is every route on an LLM listing settlement-verified?
No. Settlement-verified means a paid route with an on-chain receipt. The example links that receipt. Other routes on the listing are separate.