OPEN SOURCE · LAUNCHING SOON

Know exactly where your LLM spend goes.

Self-hosted proxy for OpenAI, Anthropic, and every other provider. Every request tagged, every dollar attributed, every month closed cleanly.

Be the first to install when v0 ships. No spam.

Thanks — you're on the list. We'll email you when v0 ships.

$187,432

Budget: $175,000 · +$12,432 (+7.1%)

Engineering — R&D

$52,180

Customer Support AI

$61,940

Marketing & Content

$28,510

Sales tooling

$18,220

Internal productivity

$14,890

What you get

Three problems, solved at the proxy.

Engineers stop stitching together provider logs. Finance stops allocating spend in spreadsheets. Founders stop guessing at gross margin per AI feature.

Attribution that doesn't require a refactor.

Drop in a virtual key per team, project, or feature. Your existing code keeps working. Spend lands in the dashboard already split by who spent it and why — no instrumentation, no SDK wrapping, no quarterly cleanup.

Monthly close reports your CFO will accept.

A real chargeback report on the first of every month. Spend by cost center, variance against budget, top movers, anomalies flagged. Exports to QuickBooks, Xero, and NetSuite. Built for the people who actually own the line item.

One proxy, every provider.

OpenAI-format and Anthropic-format endpoints. Add or swap providers without touching application code. Cache, route, and fail over between them — and watch the bill stay flat.

5-minute install

Two lines to switch your entire codebase.

Same OpenAI SDK. Same Anthropic SDK. Same code. Just point the base URL at ProxyConduit.

# Python (OpenAI SDK)
client = OpenAI(
    base_url="http://localhost:8080/v1",
    api_key="pc_live_...",
    default_headers={"x-pc-team": "engineering"}
)
    