Question 1

What is Gondola?

Accepted Answer

Gondola is a marketplace for Venice AI inference. Buyers call one OpenAI-compatible gateway and pay per request in USDC on Base, and a broker routes each call within a small price band of the cheapest healthy lender for the requested model, so the same Venice models can cost less than going direct. Those keys come from DIEM holders, who earn USDC on idle daily allowance.

Question 2

Is the API OpenAI-compatible?

Accepted Answer

Yes. Point any OpenAI client at https://api.gondola-ai.com/v1 with a Gondola key (gnd_...). The request and response shapes match OpenAI, so existing code works without changes.

Question 3

Does Gondola store my prompts?

Accepted Answer

No. Gondola never stores prompt or completion content: requests are processed in memory, forwarded to Venice for inference, and gone when the response finishes. Per-request records keep token counts and accounting metadata, never content. Venice, which runs the inference, is built around a no-logging model, and Gondola adds no logging on top. Models classified by Venice are labeled with a privacy tier: private means zero data retention on Venice-controlled infrastructure, anonymized means the request is relayed to the model provider with your identity stripped. Agents can even pay per request over x402 with no account at all.

Question 4

How are buyers charged?

Accepted Answer

Per request, in USDC. Buyers top up a balance on Base and each call debits the actual metered cost of the model used. The broker routes within a small price band of the cheapest healthy lender offering that model, so the price can fall below going direct to Venice.

Question 5

What are the fees?

Accepted Answer

Gondola charges a fee of up to 2 percent on each side. Buyers pay the posted price plus up to 2 percent (the standard rate; a strong weekly placement lowers it to as little as 0.3 percent), and lenders receive the posted price minus up to 2 percent. The broker keeps the spread. There are no subscriptions or minimums.

Question 6

How do lenders earn?

Accepted Answer

Lenders deposit a Venice INFERENCE API key and set a single price factor against Venice's list price. Gondola routes paid inference to their key and credits USDC for every request served, turning idle daily DIEM allowance into yield.

Question 7

How does the weekly fee program work?

Accepted Answer

Every week runs two competitions. Buyers rank on their own usage: place high and your platform fee drops the following week, from the standard 2% to as little as 0.3%. Lenders rank on the buyers they directly referred: placement wins routing priority the following week instead of a fee change. There is no token: both are earned by using Gondola.

The same Venice models, up to 80% cheaper.

Two sides, one gateway.

Private by construction

Same shape as OpenAI

One fee of up to 2% a side, nothing else

Step aboard.