Phaseo is building the open-source, provider-neutral gateway for AI. Discover models, route across providers, and keep your integration portable while better partnerships help move the cost of inference down.
Open source
Inspect, extend, and improve the gateway.
Provider neutral
Choose the provider that fits your work.
Lower overhead
Use efficiency and partnerships to reduce avoidable cost.
GPT-6 Astra
GPT-6 Astra
472ms
92tok/s
$10.00/$50.00
Explore benchmarks, pricing, and provider coverage to make better build decisions.
Integrate once and access hundreds of models through one OpenAI-compatible API.
Build with text, images, video, audio, moderation, embeddings, and more through one API.
GPT-6 Astra
Claude Fable 5.1
Gemini 3.8 Flash
Grok 4.6
Monitor pricing, reliability, usage, and performance for every request in one place.
Explore the broader model catalog, then filter to models with an active Gateway route before you ship.
Usage follows the model prices shown in the catalog. Today's credit purchase fee is shown openly; our mission is to reduce avoidable platform cost as provider partnerships and operating efficiency make that sustainable.
No contract, subscription, or minimum spend for model usage.
Applied when you purchase credits, with a $1 minimum. Usage follows the model prices shown in the catalog.
BYOK includes 250,000 fee-free requests each month, then a 2.5% service fee on provider-equivalent cost.
Use one gateway layer to keep model access stable as providers shift, degrade, and evolve.
Route by latency, price, geography, or your own policy layer without forcing app teams to understand every provider quirk.
Shift traffic automatically when a provider slows down, errors spike, or a model rollout creates instability.
Track model releases, deprecations, and replacements in the same platform you use to make routing decisions.
Layer scopes, keys, policies, and auditability into the gateway so reliability does not depend on tribal knowledge.
Ship with one OpenAI-compatible integration, broad provider coverage, and production-grade reliability, without lock-in or black-box infrastructure.
Send your first request in under five minutes.
Swap only the model ID to route any supported model through the same API.
View GitHub
curl https://api.phaseo.app/v1/responses \
-H "Authorization: Bearer $PHASEO_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "openai/gpt-6-astra",
"input": "Write a one-line welcome message for a new user."
}'Prefer not to integrate APIs yet?
Use Chat to test models in the browser and compare outputs with no code.
