meted dev
Run the gateway locally.
meted dev
Starts the Meted gateway on 127.0.0.1:8787. Point your OpenAI client's
baseURL at http://localhost:8787/v1.
Options
| Flag | |
|---|---|
--port <n> | Port to listen on. Default 8787 |
--host <addr> | Address to bind. Default 127.0.0.1 |
--mode <mode> | Override the configured mode for this run only |
--quiet | Do not log each request |
The banner
Meted 0.1.0 → http://127.0.0.1:8787/v1
Mode observe Compute and report the allocation. No ceiling is set on the request.
Engine baseline-heuristic 0.1.0 (open source, baseline)
Provider openai https://api.openai.com/v1
Credential OPENAI_API_KEY is set
When the credential is missing, the banner reads OPENAI_API_KEY is not set in
red, and requests are rejected with a 401 naming that variable.
The live log
ID TASK ACTUAL TARGET DIFF TOTAL ALLOCATION
a19f definition 184 38 -79% 902ms observe_mode
b91c coding 722 680 -6% 3.1s applied 960
7daa fact 91 18 -80% 410ms observe_mode truncated
A response with finish_reason: "length" is flagged truncated. An engine
failure is flagged engine:<kind>.
The log goes to stdout. Redirect it to keep it:
meted dev | tee meted.log
To collect the same fields as structured data, embed the gateway and use
onRecord.
Routes
POST /v1/chat/completions | Sized, allocated, forwarded |
ANY /v1/* | Proxied to the provider |
GET /_meted/health | Configuration and engine status |
Everything Meted decided about a request comes back on that request, as
x-meted-* headers. There is
no endpoint to read history from.
Shutdown
On Ctrl-C, meted dev stops accepting connections, lets in-flight requests
finish streaming so their token counts are measured rather than guessed, and
exits.
Security
The gateway has no authentication of its own. Anyone who can reach it can
spend your provider credits. It binds 127.0.0.1 by default. If you bind it
to a routable address with --host 0.0.0.0, put it behind something that
authenticates.