pks opencode

stable

Run OpenCode directly against a Scaleway serverless model — one command, no proxy, no config file. Defaults to GLM 5.2.

Author: Poul Kjeldager

Usage: pks opencode [ARGS] [--model <id>]

Category: infrastructure

Examples

$ pks opencode

Start OpenCode on GLM 5.2 (the default) on Scaleway

$ pks opencode --model glm-5.2

Explicitly select GLM 5.2

$ pks opencode run --format json "Reply with exactly: OK"

One-shot, non-interactive run — args pass through to the opencode CLI

pks opencode launches the OpenCode CLI directly against Scaleway's OpenAI-compatible serverless API. It is the no-proxy sibling of pks claude scaleway: where that command hosts a local Anthropic→OpenAI translation proxy for Claude Code, this one hands OpenCode a Scaleway provider and lets it call api.scaleway.ai/v1 itself.

The default model is GLM 5.2 (glm-5.2). Run pks opencode with no arguments and you are in.

Run pks your way

Use an installed pks command, run the NuGet tool directly with .NET 10, or run the npm package without installing it globally:

dotnet tool install --global pks-cli
pks scaleway init
pks opencode --model glm-5.2

Installs the permanent pks command as a .NET global tool

Prerequisites

The opencode CLI on your PATH. Install it from opencode.ai — npm install -g opencode-ai. Without it, the command exits 127 with that instruction.
A completed pks scaleway init. Stores your Scaleway secret key. Without it the command exits 1 with that instruction. If you have already authenticated for GPU or pks claude scaleway work, you are done.

Why there is no proxy and no config file

OpenCode reads its provider list from a config file (opencode.json). Editing that file by hand is friction — and writing a Scaleway secret into it is a leak. pks opencode does neither:

It reuses the secret already stored by pks scaleway init.
It builds an inline provider config as JSON and passes it through the OPENCODE_CONFIG_CONTENT environment variable — OpenCode's runtime-override mechanism, which has the highest effective priority.
It passes the Scaleway secret through PKS_SCALEWAY_API_KEY, referenced inside the inline config as {env:PKS_SCALEWAY_API_KEY}.

The result: the provider registration is process-local. Your opencode.json stays untouched, and the secret is not duplicated into OpenCode configuration or placed on the command line.

1. Start an interactive session

pks opencode

OpenCode opens against scaleway/glm-5.2. The model id passed to OpenCode is always scaleway/<model>.

2. Pick the model explicitly

pks opencode --model glm-5.2

The default is already glm-5.2, so this is equivalent — but it documents intent. Any Scaleway serverless model id is accepted verbatim, so brand-new models work before any local catalog knows about them. A scaleway/ prefix is stripped if present, so --model scaleway/glm-5.2 and --model glm-5.2 are identical.

3. Run non-interactively

pks opencode run --format json "Reply with exactly: OK"

Everything after the options is forwarded verbatim to the opencode CLI. Use run, --format, --continue, or any native OpenCode argument.

The two-command setup

If you have ever run pks scaleway init (for GPU work, for pks claude scaleway, or otherwise), you are already at step 2:

pks scaleway init              # once — stores your Scaleway secret key
pks opencode --model glm-5.2   # every time after — launches OpenCode on GLM 5.2

Options

Argument	Required	Description
`ARGS`	no	Additional arguments passed through to the `opencode` CLI (e.g. `run --format json "..."`).

Flag	Default	Description
`-m`, `--model <id>`	`glm-5.2`	Scaleway serverless model id. A `scaleway/` prefix is accepted and stripped.

Verify

pks opencode --model glm-5.2

OpenCode launches and its model selector shows scaleway/glm-5.2. Send a prompt and confirm the reply.

Troubleshooting

Symptom	Cause and fix
Exit `1` naming `pks scaleway init`	No stored Scaleway secret key. Run `pks scaleway init`.
Exit `127` naming the opencode CLI	OpenCode is not on PATH. Install it: `npm install -g opencode-ai`.
Model not found at runtime	Unrecognized ids pass through verbatim by design. Check the spelling against the Scaleway serverless catalog. The endpoint is `https://api.scaleway.ai/v1`.
Your `opencode.json` was not modified	This is intentional. The provider config is injected per-process via `OPENCODE_CONFIG_CONTENT` and never written to disk.