Model API
One connection.
Room to build.
Bring model access into the tools you are already making. Start with a straightforward HTTP request and a model of your choice.
YOUR APPLICATION / INJECT.FI
A familiar shape.
A new possibility.
Send a model ID and a list of messages. Handle the response in your own application, with your API key kept safely on the server.
- One base URL for your integration
- A model selection in each request
- Examples in cURL and JavaScript
- A quick-start guide before you connect
curl https://api.inject.fi/v1/chat/completions \
-H "Authorization: Bearer $INJECT_FI_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"YOUR_MODEL_ID","messages":[{"role":"user","content":"Hello, inject.fi."}]}'COMPUTE THAT FITS YOUR WORKFLOW
Focus on the output.
Understand the process.
Managed inference lets an application request a model response without requiring you to rent and configure a dedicated GPU first. Learn where the API fits and when direct hardware control might be useful.
Explore managed inference ↗BEFORE YOUR FIRST REQUEST
Good habits,
from the beginning.
Keep API keys out of browser code, public repositories, screenshots, and shared prompts.
Set your own application timeouts, check HTTP response codes, and review model output before using it in important decisions.
For questions about current availability or data handling for API requests, contact inject.fi before sending sensitive information.
Your next idea starts here.
Explore the models, understand the API, and take the first step with inject.fi.