# AdFreeLLM

> A public, ad-funded AI terminal. Five real models are free to use with no signup, no email, and no card — one clearly labeled sponsored card after an answer pays for the compute. A free OpenAI-compatible API is available the same way: one click, no registration.

AdFreeLLM (adfreellm.com) runs LLM inference as a public utility. There is no account system at all: the chat page and the API both work for anonymous users. The business model is stated openly on every page — advertisers sponsor one placement per answer, and that placement is what covers GPU cost instead of a subscription.

## Facts and boundaries

- Five models are exposed and each was verified end-to-end: Gemini 3.6 Flash, DeepSeek V4 Flash, Gemini 3.5 Flash Lite, GLM 5.2, Gemma 4 31B. We do not silently substitute models — if the board says DeepSeek, DeepSeek answers.
- Reliability is measured, not claimed. In sampling on 2026-08-13, GLM 5.2 returned successfully on every attempt (~725 ms average) while Gemma 4 31B failed roughly half its calls. There is no SLA; the capacity behind these models is free capacity.
- Web rate limit: about 6 messages per minute per IP. Free API key: roughly 500 short requests per day per key, two keys per IP per day, plus a site-wide daily allocation that is tuned automatically from measured upstream health.
- No account system means no stored identity, no profile, and no email tied to conversations. We do not claim privacy features we have not built.
- Ads are always labeled SPONSORED, at most one per answer. When a question involves grief, medical distress, or anything where an ad would be predatory, no ad is sent at all.
- API responses append the sponsored block as delimited text plus a structured `sponsored` field. Requests using JSON mode or tool-calling are never modified.

## Chat and models

- Terminal (pick a model and talk, no signup): https://adfreellm.com/chat
- All models: https://adfreellm.com/models
- Gemini 3.6 Flash — fast, general, the default: https://adfreellm.com/models/gemini-3.6-flash
- DeepSeek V4 Flash — reasoning and code: https://adfreellm.com/models/deepseek-v4-flash
- Gemini 3.5 Flash Lite — fastest, lightest: https://adfreellm.com/models/gemini-3.5-flash-lite
- GLM 5.2 — reasoning, strongest Chinese, most reliable in our sampling: https://adfreellm.com/models/glm-5-2
- Gemma 4 31B — open weights, least stable of the five: https://adfreellm.com/models/gemma-4-31b

## Free API

- Documentation, live key button, and code samples: https://adfreellm.com/api
- Base URL: https://api.adfreellm.com/v1 (OpenAI-compatible `/chat/completions`)
- Mint a key with no signup: `POST https://api.adfreellm.com/v1/keys`
- Current daily availability: `GET https://api.adfreellm.com/v1/keys/status`

## Pay-per-call research APIs for agents

- Catalog and integration guide: https://adfreellm.com/agent-services/
- Five bounded JSON services: Google Search ($0.01), web crawl to Markdown ($0.04), Google Maps businesses ($0.03), Instagram research ($0.02), and YouTube research ($0.03).
- Payments use x402 v2 with USDC on Base. An unpaid request returns HTTP 402 with the exact payment requirement; an x402 client signs, pays, and retries automatically.
- These research services are separate from the free OpenAI-compatible LLM API above. They expose public-data research primitives, not raw LLM inference.

## Pay-per-token LLM APIs

- Paid inference catalog: https://adfreellm.com/llm-services/
- Machine-readable service directory: https://adfreellm.com/llm-services/catalog.json
- Eight OpenAI-compatible models are available through two independent payment rails: Halo and AntSeed.
- Halo publishes one flat rate of 1 USDC per million total tokens across all eight models. AntSeed publishes model-specific input, output, and cached-input rates in signed provider metadata.
- Both payment rails settle in USDC on Base. These paid routes are separate from the anonymous, ad-funded free API documented above.

## How it is funded

- The open ledger — where the ad appears, how the money covers compute, why that removes signups: https://adfreellm.com/how-advertisers-pay
- Buy a placement (rates, targeting, house rules): https://adfreellm.com/advertise

## Comparisons

- vs ChatGPT free tier: https://adfreellm.com/compare/adfreellm-vs-chatgpt-free
- vs DeepSeek official: https://adfreellm.com/compare/adfreellm-vs-deepseek
- vs Gemini official: https://adfreellm.com/compare/adfreellm-vs-gemini

## Optional

- FAQ (is it free, do I need an account, rate limits, where ads appear, API): https://adfreellm.com/faq
- Sitemap: https://adfreellm.com/sitemap.xml