How it works
Drop in. Every model, live.
Hicap is a drop-in replacement for the OpenAI API. Swap your base URL, keep your client, and reach every major model. Here is the whole integration.
From base URL to dashboard.
Four steps, all in the client you already run: swap the endpoint, let the gateway route each call, tag your requests, and watch usage land in the dashboard.
Swap your base URL
Hicap is a drop-in replacement for the OpenAI API. Point your existing SDK, CLI tool, or extension at our endpoint — no new client to learn.
- Works with the OpenAI client you already use
- Chat, streaming, tool calls, and embeddings, unchanged
- One line of config, integrated in minutes
1curl https://api.hicap.ai/v1/chat/completions \2 -H "api-key: $HICAP_API_KEY" \3 -H "Content-Type: application/json" \4 -d '{5 "model": "gpt-5.4",6 "messages": [7 { "role": "user", "content": "Hello" }8 ]9 }'
Requests flow through the gateway
Every call hits the Hicap gateway, which places it on your reserved capacity and overflows to on-demand when a model runs hot — one endpoint in front of every provider.
- One endpoint reaches every major provider
- Reserved capacity absorbs your baseline
- Overflow spills to on-demand automatically
Tag requests for attribution
Attach an x-hicap-tags header to each request. Hicap records the tags on every log line and rolls them into spend attribution — set once in your client.
- Tag by feature, team, app, or any dimension
- Attribution reads by tag, not by invoice
- No separate tracking pipeline to maintain
1curl https://api.hicap.ai/v1/chat/completions \2 -H "api-key: $HICAP_API_KEY" \3 -H "Content-Type: application/json" \4 -H 'x-hicap-tags: {"feature": "product-recommendations", "team": "growth", "app": "web"}' \5 -d '{6 "model": "gpt-5.4",7 "messages": [8 { "role": "user", "content": "Recommend a product" }9 ]10 }'
See every request in the dashboard
Tokens, cost, and requested-vs-served model for every call — across your dev tools and production apps, in one place.
- Usage by model, app, and team
- Dev tooling vs production at a glance
- Spot your most expensive models
Coding tools
Dev tooling with BYOK configuration.
$246.45
28.6M tokens
Your app
Production workload with reserved capacity.
$1,704.00
10,500 TPM + 66.4M
Built for production.
The AI control plane behind your stack — one AI gateway for LLM and voice, with observability, cost control, and governance built in.
Drop-in compatible
Works with the OpenAI client you already run — chat, streaming, tool calls, and embeddings, unchanged.
Every provider, one key
Reach every major provider through a single endpoint. Change a model string to switch providers.
No data retention
We never store your prompts or completions. Requests pass through the gateway and are never retained.
Real-time analytics
Usage, cost, and per-model performance in the dashboard as requests happen.
Go live
Swap your base URL and ship.
Create an account, change one line, and every model is live. Setup takes minutes. For routing, savings, and reliability in depth, see LLM Gateway.