Inference that lets youship a feature
Run open models on an OpenAI-compatible API, serverless or dedicated, on the same key.
curl https://tokens.flex.ai/v1/chat/completions \
-H "Authorization: Bearer $FLEXAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "Meta-Llama-3.1-8B-Instruct-FP8",
"messages": [{"role": "user", "content": "Hello from FlexAI"}]
}'Results in production
Real outcomes from teams running inference on FlexAI.
Built for agent loops
The inference API gives your agent the primitives; the Agent SDK adds the run-level controls.
From the inference API
- ✓ Tool calls
- ✓ Structured output
- ✓ Streaming
- ✓ Model fallback
With the Agent SDK in trial
- ✓ Eval harnesses
- ✓ Approval logic
- ✓ Per-run observability
One API shape,serverless or dedicated
Start on shared per-token endpoints. When steady volume justifies your own GPUs, dedicated endpoints serve the same OpenAI-compatible API. Switch the model id, keep your code.
import os
from openai import OpenAI
client = OpenAI(
base_url="https://tokens.flex.ai/v1",
api_key=os.environ["FLEXAI_API_KEY"],
)
response = client.chat.completions.create(
model="Meta-Llama-3.1-8B-Instruct-FP8",
messages=[{"role": "user", "content": "Hello"}],
)Production visibility without guesswork
Understand what happened without digging through tools and logs.
Live usage metrics
Latency, tokens, GPU, region, cache. In one view.
Cost lens
What is expensive, what is waste, what to fix next.
AutoScaling
Scale Fractional or Full GPUs based on traffic.
Production ready blueprints
Ship inference like it isa feature, not a project
Built for teams shipping real inference to real users. Sizing a workload? Compare costs in the GPU savings calculator or see dedicated endpoints for dedicated capacity.
Serverless
Per-token billing at competitive rates.
Dedicated
Your own endpoints on owned GPUs, billed per second.
Same API
Move between them without a rewrite.
Proof
Scaling inference across clouds
“The ability to manage our compute resources across multiple cloud providers through a unified interface is a game-changer.”
Pixelcut
Read the case study<24h
Time to first production inference deploy
Frequently Asked Questions
More managed AI services: fine-tuning, training, and the platform overview.