dsh-devin2api
Devin2API model provider for DeepSeek Harness
설치
$
npx -p @deepseek-ai/dsh dsh plugin --profile web add github:Green-hats/dsh-devin2apiConfiguration
Everything lives under providers.devin2api:
| Field | Default | Notes |
|---|---|---|
displayName | Devin2API | Provider name in the selector |
baseURL | http://127.0.0.1:18089/v1 | Gateway root including /v1 |
apiKey / apiKeyEnv | '' / DEVIN2API_API_KEY | Direct key (redacted) or env var read per request |
enabled | true | Mount-time gate; changing it needs a reload |
models | [] | Empty = all. Non-empty allowlist of ids or { id, name } |
modelVisibility | {} | Per-model switch; false hides a model |
images | true | Declare image input for vision models |
contextWindow / maxTokens | 262144 / 128000 | Limits reported to DSH (not in /v1/models) |
refreshMinutes | 5 | Catalog cache TTL (minimum 1) |
requestTimeoutMs | 300000 | Per-request timeout including the stream |
imagePixelBudget / imageMaxBytes | 4194304 / 1048576 | Resize target per image |
maxRequestImageBytes | 20971520 | Total base64 characters per request |
All fields except enabled are volatile: settings writes apply live. DEVIN2API_BASE_URL is reported to the Models
page as an endpoint hint.