Prototypes that need a backend
You wanted to test a prompt. You ended up configuring a project.
Playground
Prompt, run, see the price
The fastest way to find out if an idea works. Pick a model or let the router choose, write a prompt, and see the response alongside its tokens, latency and cost. No code required.
No code
test ideas in the browser
Real numbers
tokens, latency and price per run
Straight to production
same request, same endpoint
The problem
What is slowing you down today
No idea what it will cost
The demo looks great, but nobody knows the price per call at scale.
Playgrounds that lie
A tool that behaves differently from the real API wastes everyone's time.
Capabilities
Everything you need, nothing you don't
Every catalog model
Pick a named model, or leave it unset and let the router choose.
Temperature and max tokens
Tune the same parameters you will send from code.
Price on every run
Input tokens, output tokens, latency and price shown with each response.
Same endpoint as production
The playground runs through the real inference path, so what you see is what you ship.
Cache visibility
Run the same prompt twice and see the cached result come back unbilled.
Lives in your history
Playground runs appear in request history alongside API traffic.
Use cases
What teams run on it
01
Prompt design
Iterate on wording until the output is right.
02
Model comparisons
Same prompt, different models, side by side in your history.
03
Cost estimates
Measure a typical request before you plan a rollout.
04
Stakeholder demos
Show a colleague a working idea in seconds.
How it works
From sign-up to production
- 1
Open the console
Sign in and choose Playground.
- 2
Write and run
Choose a model or let the router pick, set temperature and max tokens, and run.
- 3
Copy to code
Send the same prompt, model_id and parameters to POST /api/ainf/inference.
Control
Built so you stay in charge
Keys you control
API keys are shown once at creation and stored only as a hash. Revoke any key instantly from Settings.
Spend you can see
Every request carries its own receipt: tokens in, tokens out, latency and price. Nothing is billed that you cannot trace.
Limits you set
Per-account rate limits and a free monthly allowance keep experiments safe before they become production.
Data you decide on
You choose exactly what you send. Repeat prompts can be served from your cache, and cache hits are never billed.
For developers
One endpoint. Every model.
Send a prompt and a capability, or name a model. One bearer key, one small JSON body, and a full receipt on every response.
curl https://YOUR-DOMAIN/api/ainf/inference \
-H "Authorization: Bearer $AI_INFERENCE_KEY" \
-H "Content-Type: application/json" \
-H "Idempotency-Key: order-1042" \
-d '{"prompt":"Summarise this ticket in one line","capability":"chat"}'Questions, answered
Are playground runs billed?
They are metered like any other call and count toward your free monthly allowance first. Cached repeats are not billed.
Do I need an API key?
No. The playground uses your signed-in console session.
Is the playground the same as the API?
Yes. It calls the same inference endpoint with the same routing, caching and pricing.
Can I see past runs?
Every run appears in request history with its model, tokens, latency and price.
Your next idea is one prompt away
Open the playground and run it free.