OpenAI-compatible model gateway

One API for models, routing, usage, and settlement.

GotoAI connects developers and teams to multiple model providers through one OpenAI-compatible API — with provider routing, usage controls, and a request-level billing ledger. Enterprise governance is on the roadmap.

Start building View docs
OpenAI SDK compatibleRequest-level ledger Provider routing & fallbackMetadata-only observability Enterprise controls roadmap
Compatible with the OpenAI SDK Routes across providers with fallback Reconciles every request in a ledger Stores metadata only — never prompts BYOK & audit on the roadmap
The problem

Direct provider integration is fragile.

Wiring each app straight to one provider means single points of failure, no cost control, scattered keys, and no auditable record of what was spent. As you add providers and teams, that complexity compounds.

Unified API

One OpenAI-compatible endpoint

Keep your code. Change base_url and key, and route every model through one gateway.

base_url = "{{api_base}}"
model    = "{{default_model}}"
Supply Mesh

Multi-provider routing & fallback

Score candidates by cost, latency, quality and health — then fail over to the next eligible provider automatically.

ProviderOut /1MHealth
openrouter$0.60closed
openai$0.60closed
anthropic$15.0closed
Clearinghouse

Request-level clearing & ledger

Reserve before execution, finalize by real tokens, and reconcile drift — so AI spend never surprises you.

EntryAmountBalance
recharge+$10.00$10.00
reserve−$0.0009$9.9991
release+$0.0009$10.00
charge−$0.0000071$9.99999

Console

Create keys, debug in the playground, watch usage, inspect the ledger, and manage supply and risk — all from one control panel.

Console →

Enterprise roadmap

BYOK, audit export, org-scoped administration, RBAC and SSO are gated roadmap modules. Talk to us for early access.

Security & trust →
How it works

From request to settled ledger

1
Request
Your app sends an OpenAI-compatible request.
2
Route
The router selects a provider by policy and health.
3
Reserve
Risk gates check limits and freeze the max cost.
4
Execute
The provider runs; fallback covers failures.
5
Finalize
The ledger charges real tokens and reconciles.

OpenAI migration

You already use the OpenAI SDK but want multi-provider supply and unified governance.

Routing & fallback

One provider's outage, rate limit or price spike shouldn't take down your app.

Team cost governance

Org keys, per-key limits, and a ledger that makes spend auditable.

Quickstart

Your first call in minutes

Set your key and base URL, then call any model through one endpoint. No new SDK to learn.

Read the quickstart
from openai import OpenAI

client = OpenAI(
    api_key=os.environ["GOTOAI_API_KEY"],
    base_url="{{api_base}}",
)
print(client.chat.completions.create(
    model="{{default_model}}",
    messages=[{"role":"user","content":"Hello"}],
    max_tokens={{max_tokens}},
).choices[0].message.content)
curl {{api_base}}/chat/completions \
  -H "Authorization: Bearer $GOTOAI_API_KEY" \
  -d '{"model":"{{default_model}}","messages":[{"role":"user","content":"Hello"}],"max_tokens":{{max_tokens}}}'
Pricing

Platform control, separate from model usage

Upstream token cost is billed separately from the GotoAI platform fee. BYOK is an Enterprise roadmap option.

…
See full pricing →
Security & trust

Built for control, not just calls

Concrete guarantees, not slogans.

Read security & trust
✓
API keys stored as hashes
Plaintext key shown once at creation.
✓
Provider source is visible
Response header shows the upstream provider.
✓
Metadata-only usage events
Prompts and responses are never stored.
✓
Reconcilable request ledger
Reserve → finalize, with drift checks.

Start routing through one API.

Create a key, call any model, and see the ledger settle each request.

Start building Contact sales