Quickstart
Point any OpenAI-compatible client at one base URL. Keep your keys (BYOK). Connect MCP hosts from the integrations board when you want agents in the loop.
Paths
- Seat: /billing/intermediate — Checkout issues your Metrecept key.
- Try without an IDE: /workbench Agent Shell, or /demo for the 60s miss→HIT proof.
- SDK base URL:
https://api.withohm.dev/v1(local edge:http://localhost:8081/v1). - BYOK: send your provider key as
X-Ohm-Upstream-Key(X-Metrecept-Upstream-Keyis dual-accepted). Authorization stayssk-at-…. - Govern: /org — cost centers, ledger statement, policy.
- Connect tools: Integrations — Cursor, Claude Code, VS Code, Windsurf, Zed.
- Enterprise: /billing/enterprise.
Python
Identical second call → Redis cache hit (x-at-cache: HIT).
from openai import OpenAI
client = OpenAI(
api_key="sk-at-YOUR_OHM_KEY",
base_url="https://api.withohm.dev/v1",
default_headers={"X-Ohm-Upstream-Key": "sk-proj-..."}, # BYOK on miss
)
completion = client.chat.completions.create(
model="gpt-4o-mini",
messages=[{"role": "user", "content": "hi"}],
)
curl
Self-proof runbook: same commands as Trust — verify it yourself, scripted for CI.
curl -s https://api.withohm.dev/v1/chat/completions \
-H "Authorization: Bearer sk-at-YOUR_OHM_KEY" \
-H "X-Ohm-Upstream-Key: $OPENAI_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"mock","messages":[{"role":"user","content":"ohm-self-proof-v1"}]}'