Getting started

Fast Inference gives you one OpenAI-compatible API for fast hosted models. A new account starts with a personal team that holds one project, and every API key, request, and usage record in fast-web is scoped to the first project of the team selected in the dashboard sidebar.

#Send your first request

1. Create your account

   Create an account at /register, or sign in with an existing account. Fast-web creates your first project automatically and opens the onboarding flow.

2. Create a project API key

   Open API keys, choose Create API key, and give the key a recognizable name. The dashboard shows the raw key once.

3. Export the key

   Set the key in your current shell. Do not commit it to source control.

   bashCopy

   export INFERENCE_API_KEY="inf_sk_..."

4. Call the API

   Send an OpenAI-compatible chat completion. The model ID is the same value shown in the model catalog.

   curltypescriptpython

   Copy

   curl https://api.inference.net/v1/chat/completions \
     -H "Authorization: Bearer $INFERENCE_API_KEY" \
     -H "Content-Type: application/json" \
     -d '{
       "model": "glm-5.2",
       "messages": [
         { "role": "user", "content": "Explain speculative decoding in one sentence." }
       ]
     }'

   A successful response contains an assistant message under choices[0].message.content and token counts under usage.

5. Verify usage

   Open Usage. The request appears in the spend graph and model table after ingestion completes. Hover a bar to inspect spend and tokens for each model.

#Use an SDK

Any OpenAI-compatible client can use the same endpoint and key.

import OpenAI from "openai"

const client = new OpenAI({
baseURL: "https://api.inference.net/v1",
apiKey: process.env.INFERENCE_API_KEY,
})

const response = await client.chat.completions.create({
model: "glm-5.2",
messages: [{ role: "user", content: "Hello from Fast Inference" }],
})

console.log(response.choices[0].message.content)

#Understand project scope

Fast-web operates on the first project of the team selected in the dashboard sidebar. The selection defaults to your personal team and is remembered in this browser; it is independent of the team shown on the main Inference dashboard. The fast CLI asks you to select a team and project during interactive login; select that team in the sidebar to see what the CLI created. In a non-interactive terminal, it falls back to your personal team's first project. Keys created in the dashboard are granted read and write access to that project. Usage and analytics are loaded for the same scope, so a request made with the key should appear in the dashboard without choosing a project manually.

#Troubleshooting

Next, install the CLI or connect a coding agent from the integration guides.