Every model,
one key.Hold DEF,
receive INFERENCE.
Run 449 models from Claude Code, Codex or any SDK, paid in USDC Sepolia testnet. DEF trading fees buy INFERENCE for holders.
- 449 models
- No Deference markup
- $2.50 free credit to start
How it works
You pay in USDC. DEF holders receive INFERENCE.
Your USDC goes into an on-chain reserve that mints INFERENCE, and activating it gives you credit for any model. Trading fees from DEF mint more for holders to spend, keep or sell.
You
- Your USDC
- INFERENCE from the reserve
- Activated as credit
- Any model
DEF holders
- DEF trades
- Fees in USDC
- Buys INFERENCE from the reserve
- Split by time‑weighted balance
Reserve
SepoliaINFERENCE trades on Uniswap at the market price. The reserve does not buy it back.
Reserve and usage statsConnect
Every model at OpenRouter's price. Change two lines to switch.
Point Claude Code, Codex, Cursor or any OpenAI or Anthropic SDK at one base URL. OpenRouter's price plus OpenRouter's 5% platform fee. Deference adds no markup.
~/.zshrc
export ANTHROPIC_BASE_URL="https://deference.si"export ANTHROPIC_AUTH_TOKEN="$DEFERENCE_API_KEY"export CLAUDE_CODE_ENABLE_GATEWAY_MODEL_DISCOVERY="1"
~/.codex/config.toml
model = "openai/gpt-5.3-codex"model_provider = "deference"web_search = "disabled"[model_providers.deference]name = "Deference"base_url = "https://deference.si/v1"wire_api = "responses"env_key = "DEFERENCE_API_KEY"
Settings › Models › API Keys
opencode.json
{"provider": {"deference": {"npm": "@ai-sdk/openai-compatible","name": "Deference","options": {"baseURL": "https://deference.si/v1","apiKey": "{env:DEFERENCE_API_KEY}"}}},"model": "deference/anthropic/claude-sonnet-5.5"}
app.py
import osfrom openai import OpenAIclient = OpenAI(base_url="https://deference.si/v1",api_key=os.environ["DEFERENCE_API_KEY"],)
One address, two API shapes
- Chat Completions, Responses and Messages
- Streaming
- Tool calls and structured outputs
- Docs your agent can read
Rewards
Hold DEF. Its trading fees reach you as INFERENCE.
Each hour with trades, DEF's USDC fees buy INFERENCE and split it by balance and time held. It lands in your wallet to spend on any model, keep or sell.
Per 1,000 USDC traded: 5 INFERENCE to holders
- Nothing to stake or lock
- No minimum balance
- Your share follows time held
- Sent to your wallet from 1 INFERENCE
- Spend as credit on any model
- Every payout’s data is public
Control
A limit on every key. A cost on every request.
Give each tool, script or teammate its own key with a credit limit and an expiry. See spend by key and model, and each request's tokens, latency and cost.
Example: one Claude Sonnet 5.5 request on the production key, shown in Keys, Usage and Activity, final cost $0.0062.
Example data
Questions
Read the docsCan I use it today?
Yes, on the Sepolia testnet. Create a key, use up to $2.50 of free credit on open-weight models, and add credit with test USDC.
How is a request priced?
OpenRouter's price plus OpenRouter's 5% platform fee. Deference adds no markup. Each request holds its worst-case cost first, then settles at the metered cost.
How is this different from OpenRouter?
The same models at OpenRouter's price. You pay in USDC, the reserve behind INFERENCE is readable on chain, and DEF holders receive INFERENCE from trading fees.
Why hold DEF?
Each hour with trades, its USDC fees buy INFERENCE for holders by balance and time held. Activate it into credit, keep it or sell it on Uniswap. Amounts follow trading volume, and nothing is distributed without trades.
Can I get my money back out?
Credit cannot be withdrawn or sent. INFERENCE you have not activated can be sold on Uniswap at the market price; the reserve does not buy it back.
Do you store my prompts?
No. Deference keeps request metadata, such as model, tokens, latency and cost, not the content. OpenRouter and the model providers have their own policies.
Which tools work with Deference?
Claude Code, Codex, Cursor (beta), OpenCode, Cline, Zed, Continue, Goose and more, each with a setup guide. Gemini CLI does not; its request format is not supported.
Start with up to $2.50 of free credit.
Verify an email, create a key and make your first request. Each step adds free credit for open-weight models.
Example: the get-started checklist after the first three steps, with $1.00 of free credit added.