All the free LLMs.
One API.

Quickstart

Point your SDK here.

Anything that speaks the OpenAI Chat Completions API works. Set the base_url, paste your key, done. Get a key →

curl https://dreamprompting.com/api/v1/chat/completions \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "auto",
    "messages": [{"role":"user","content":"Hello!"}]
  }'

from openai import OpenAI

client = OpenAI(
    base_url="https://dreamprompting.com/api/v1",
    api_key="YOUR_API_KEY",
)
r = client.chat.completions.create(
    model="auto",
    messages=[{"role": "user", "content": "Hi"}],
)
print(r.choices[0].message.content)

const r = await fetch("https://dreamprompting.com/api/v1/chat/completions", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    "Authorization": "Bearer YOUR_API_KEY",
  },
  body: JSON.stringify({
    model: "auto",
    messages: [{ role: "user", content: "Hi" }],
  }),
});
const d = await r.json();
console.log(d.choices[0].message.content);

Pin a model with "model": "groq/openai/gpt-oss-120b", or use "auto" to let the router pick and fail over. Streaming works: pass "stream": true or the SDK's stream=True. List everything at /api/v1/models.

Routing

The route table

Requests try live providers in order and fail over on errors or rate limits. Free tiers first, trial credits next, keyless ch.at as the always-on last resort.

Models

Popular models at $0

DreamPrompting is a free, OpenAI-compatible LLM API and a drop-in free alternative to the ChatGPT and OpenAI API. Because it speaks the standard chat completions format, you can call popular open and GPT-class models for free: chat, coding, summarization, extraction and content generation, all from one endpoint with one free key.

Browse and test everything on the models page, or see every AI model ever released.

Free tool, sign in to scan

See what is hiding in your code

Zero-width characters, bidirectional overrides that make source display in one order and compile in another, and Unicode tag characters carrying a payload that renders as nothing. Every removal is reported by codepoint and line, so you can check the change instead of trusting it.

Free tools, no signup

Work out the cost before you pay it

Four calculators built on real data, not rules of thumb: a live BPE tokenizer, plus architecture figures for 63 open-weight models and 34 cards. Nothing to install, no account needed.

Or run it on your own GPU