lee/puter
1
0 Fork 0
mirror of https://github.com/HeyPuter/puter.git synced 2026-10-17 09:09:43 +00:00
Code Issues Actions 8 Packages Projects Releases Wiki Activity
Files
facb84474733eaa833cc6c453e5eb5cd784272be
puter/config.template.jsonc
T
Nariman Jelveh 19cd5f4476 feat: add BytePlus ModelArk providers (chat, image, video) (#3498) 
* feat: add BytePlus ModelArk chat provider

Adds BytePlus ModelArk as a provider for the puter-chat-completion
driver, following the MiniMax/ZAI providers as reference per
doc/contributing-apis.md.

- OpenAI-compatible endpoint at ark.ap-southeast.bytepluses.com/api/v3
  (apiBaseUrl config selects the region)
- Static catalog of 16 chat models (Seed 2.x/1.x incl. vision, GLM,
  DeepSeek, GPT-OSS) with limits and per-token pricing from the
  official docs
- Passes Ark's thinking/response_format/stop params through custom;
  normalizes reasoning_content to reasoning
- Bare deepseek-v4-* names stay with the first-party DeepSeek
  provider; BytePlus only claims prefixed aliases
- Offline unit tests (mocked SDK against a real test server) plus an
  env-gated integration test

* feat: add BytePlus image and video providers

Extends the BytePlus ModelArk integration to the puter-image-generation
and puter-video-generation drivers, reusing the same services.byteplus
API key and regional apiBaseUrl as the chat provider.

Image (Seedream/SeedEdit via OpenAI-compatible /images/generations):
- dola-seedream-5-0-pro (pixel-tier pricing + billed input images from
  the 2nd on), seedream-5-0-lite, 4-5, 4-0, and seededit-3-0-i2i
- quality tiers 1K/1.5K/2K; aspect ratios resolve to Ark's documented
  pixel sizes; explicit WxH passes through with Ark's bounds enforced

Video (Seedance via Ark's async /contents/generations/tasks + polling):
- Seedance 2.0 / 2.0 Fast / 2.0 Mini / 1.5 Pro / 1.0 Pro / 1.0 Pro Fast
  (2.5 is priced but its API isn't live yet, so it's excluded)
- per-video-token billing from usage.completion_tokens, with per-second
  estimates feeding the credit cap; audio vs silent rates for 1.5 Pro
- first/last frame and reference-image inputs; generate_audio param
  added to IGenerateVideoParams

Pricing and capabilities hardcoded from the official docs (ModelArk
pages 1544106, 1330310, 1520757, 1521309, 1541523). Offline unit tests
mock the SDK / global fetch; integration tests are env-gated on
PUTER_TEST_AI_BYTEPLUS_API_KEY.

* fix: correct BytePlus catalogs and validation against the live API

Verified the three BytePlus providers against ModelArk with a real key;
these are the mismatches that surfaced.

- Drop seededit-3-0-i2i-250628. Ark reports it as Shutdown and every
  request 404s. Its now-unreachable image-to-image branches in the
  provider go with it.
- seedream-4-5 and the 5.0 series enforce a 3,686,400 pixel minimum, so
  they only accept the 2K tier. Mark them 2k-only and snap an
  unsupported tier up to the nearest allowed one, which also keeps the
  aspect-ratio table from mapping to a sub-minimum size.
- glm-4-7 has a 204,800 token context, not 256K.
- Guard the actor in the image provider like the video provider does.
- Round a sub-minimum video duration up to the shortest supported clip
  instead of reporting it as insufficient funds.
- Gate video resolution on the model's own dimensions; the dims table is
  shared across a family and accepts more than any one model does.

* Tighten BytePlus AI provider handling

Extract shared reasoning-content normalization for OpenAI-style chat providers, and harden BytePlus image/video behavior. This updates image tier and size validation, normalizes aspect ratios and input image refs, prevents mismatched BytePlus key/base URL fallback config, makes video resolution matching case-insensitive, and rejects excess reference images instead of silently truncating them. Tests were expanded to cover the new BytePlus request and validation paths.
2026-08-15 21:08:17 -07:00

474 lines
21 KiB
JSON
Raw Blame History

{
// Comprehensive template — every key the OSS backend or shipped (non-prod)
// extensions read. Copy to `config.json` and trim what you don't need;
// unset keys fall back to documented defaults. See
// `src/backend/types.ts` for the per-field source of truth.
//
// Each setting lives at exactly one canonical key — there are no fallback
// aliases. Values shown are illustrative, not production secrets.
//
// Keys consumed only by closed-source / hosted-prod extensions (clickhouse,
// cacheUpdateHandler, pages, prodMeteringAndBilling, …) are intentionally
// omitted.
// ── Environment / identity ──────────────────────────────────────────
"config_name": "template",
// `dev` opens a browser on boot, skips blocked-email checks, and runs the
// dev-time webpack watcher; `prod` serves pre-built bundles.
"env": "dev",
// Console output format. `json` replaces the global console so every call
// emits one structured JSON line (level, timestamp, msg, and the active
// request's trace id) — one event per call, so a line-oriented log
// collector can't split stack traces across events and level filtering
// works. Unset (the default) leaves console output human-readable.
"log_format": "text",
"version": "0.0.0",
// Stable identity for this server node — used by pager alerts and
// graceful-shutdown coordination.
"serverId": "node-1",
// ── Networking / URLs ───────────────────────────────────────────────
// Port Puter listens on internally.
"port": 4100,
// Externally-visible port (set this when behind a reverse proxy on 80/443).
"pub_port": 4100,
"protocol": "http",
"domain": "puter.localhost",
// Fully-qualified externally-visible URL. Computed from protocol/domain/
// pub_port if unset.
"origin": "http://puter.localhost:4100",
// Public base URL for the API subdomain (used to build signed URLs and
// surfaced to the client by the `installedApps` and `whoami` extensions).
"api_base_url": "http://api.puter.localhost:4100",
// Subdomains Puter routes on. Wildcard DNS (`*.<domain>`) must point at
// this server for site/app hosting to work.
"static_hosting_domain": "site.puter.localhost",
"static_hosting_domain_alt": "host.puter.localhost",
"private_app_hosting_domain": "app.puter.localhost",
"private_app_hosting_domain_alt": "dev.puter.localhost",
// Host-header / domain handling. Defaults below are the dev-friendly
// settings; tighten for any public install.
"allow_all_host_values": true,
"allow_no_host_header": true,
"allow_nipio_domains": false,
"custom_domains_enabled": false,
"enable_ip_validation": false,
// Express `trust proxy` setting — set to the number of reverse-proxy
// hops in front of the server (1 = nginx/Cloudflare, 2 = CF→ALB→app),
// or to a CIDR / IP / list. `false` (safe default) makes `req.ip` the
// direct socket peer. NEVER set to `true` in prod — it trusts every hop
// and makes X-Forwarded-For forgeable.
"trust_proxy": false,
"no_browser_launch": false,
// ── Dev watcher (devWatcher extension) ──────────────────────────────
// Rebuilds GUI + puter.js on file changes when running from source.
// Ignored when `env: "prod"` unless `devwatch.enabled: true`.
"no_devwatch": false,
"devwatch": {
// Delay after watcher startup before boot continues. Lets webpack
// emit its first build so the homepage doesn't 404 on bundle.min.js.
"ready_delay_ms": 5000
},
// ── Auth / session ──────────────────────────────────────────────────
// ALWAYS replace these for any public install — `openssl rand -hex 64`.
"jwt_secret_v2": "change-me",
"url_signature_secret": "change-me",
"cookie_name": "puter_auth_token",
"min_pass_length": 6,
// When true, anonymous users must log in instead of creating temp or
// permanent accounts.
"disable_user_signup": false,
"allow_system_login": false,
"strict_email_verification_required": false,
"captcha": {
"enabled": false,
"difficulty": "medium"
},
"oidc": {
"providers": {
// Google uses OIDC discovery — only ids are required.
"google": {
"client_id": "",
"client_secret": "",
"scopes": "openid email profile"
},
// Custom OIDC providers must also supply the three endpoints.
"custom-oidc": {
"client_id": "",
"client_secret": "",
"authorization_endpoint": "",
"token_endpoint": "",
"userinfo_endpoint": "",
"scopes": "openid email profile"
}
}
},
// ── Groups / provisioning ───────────────────────────────────────────
// UIDs of the persistent groups new users are auto-enrolled in.
"default_user_group": "78b1b1dd-c959-44d2-b02c-8735671f9997",
"default_temp_group": "b7220104-7905-4985-b996-649fdcdb3c8f",
// When true, ACL grants read/list on `/<user>/Public` to any actor.
"enable_public_folders": true,
// ── Storage / S3 ────────────────────────────────────────────────────
"s3": {
// Local fauxqs (in-process S3-compatible) — used in dev and the
// bundled-defaults Docker mode. Files land under `dataDir`.
"localConfig": {
"inMemory": false,
"host": "127.0.0.1",
"port": 4566,
"dataDir": "volatile/runtime/fauxqs-data",
"s3StorageDir": "volatile/runtime/fauxqs-s3-data"
},
// For real / external S3, replace `localConfig` above with `s3Config`.
"_remote_example": {
"s3Config": {
"useCredentialChain": false,
"endpoint": "https://s3.example.com",
// Endpoint used in presigned URLs handed to the browser. Set
// this when the server-side endpoint isn't reachable from the
// browser (e.g. docker-internal `http://s3:9000`).
"publicEndpoint": "",
"accessKeyId": "",
"secretAccessKey": "",
"region": "us-west-2",
// Set true for RustFS / MinIO / fauxqs (path-style URLs).
// Real AWS S3 wants virtual-hosted — leave unset / false.
"forcePathStyle": false
}
}
},
"s3_bucket": "puter-local",
"s3_region": "us-west-2",
"region": "us-west-2",
// Default per-user storage cap (bytes). 100 MB.
"storage_capacity": 104857600,
"is_storage_limited": false,
"available_device_storage": 0,
// ── Thumbnails (thumbnails extension) ───────────────────────────────
// Optional dedicated S3-compatible bucket for generated thumbnails.
// When unset (or `endpoint` empty), the extension falls back to the
// main S3 client / bucket above.
"thumbnailStore": {
"name": "puter-local",
"endpoint": "",
"credentials": {
"accessKeyId": "",
"secretAccessKey": ""
}
},
// ── Database ────────────────────────────────────────────────────────
"database": {
// `sqlite` for single-node/dev, `mysql` for MariaDB/MySQL,
// or `postgres` for PostgreSQL.
//
// NOTE: `postgres` support is a community contribution and is not
// exercised by Puter.com production. It boots, passes its own
// integration tests against pgmock, and runs the common user/app/
// fsentry/session/permission/OIDC flows — but less-traveled code
// paths may surface MySQL/SQLite-isms that haven't been ported yet.
// Expect rough edges and please file issues if you hit one. For
// production self-hosting today, `mysql` (MariaDB) and `sqlite` are
// the supported defaults.
"engine": "sqlite",
// sqlite — file path on disk
"path": "volatile/runtime/puter-database.sqlite",
"targetVersion": 0,
// mysql/postgres — connection details. PostgreSQL defaults to port
// 5432 when `engine` is `postgres`; MySQL/MariaDB normally use 3306.
"host": "",
"port": 3306,
"user": "",
"password": "",
"database": "",
// postgres may also use a URL instead of host/user/password fields:
// "connectionString": "postgres://puter:secret@localhost:5432/puter",
// Optional read-replica pool. Reads route here when populated.
"replica": {
"host": "",
"port": 3306,
"user": "",
"password": "",
"database": ""
}
// mysql/postgres self-host bootstrap: set `migrationPaths` to apply the
// bundled schema on first boot. Idempotent — safe to leave on.
// "migrationPaths": ["./src/backend/clients/database/migrations/mysql"]
// "migrationPaths": ["./src/backend/clients/database/migrations/postgres"]
},
// ── DynamoDB (KV store) ─────────────────────────────────────────────
"dynamo": {
// Local emulator (dynamodb-local) endpoint. Drop this field for real
// AWS DynamoDB.
"endpoint": "http://localhost:8000",
// Set true when pointing at a local emulator so Puter creates the KV
// table on boot. NEVER set against real AWS — provision via IaC.
// "bootstrapTables": true,
"path": "",
// Credentials. NOTE: snake_case here, unlike `s3.s3Config` below.
// For dynamodb-local, any non-empty values work.
"aws": {
"access_key": "",
"secret_key": "",
"region": "us-west-2"
}
},
// ── Redis / Valkey (cache + cross-node rate limit) ──────────────────
"redis": {
// True → in-process redis-mock (dev / single-node).
"useMock": true,
// Cluster nodes for ioredis. For a single Valkey/Redis container,
// run it in cluster mode (one node, all slots).
"startupNodes": [
{
"host": "127.0.0.1",
"port": 7000
}
]
// Defaults to true (matches prod ElastiCache). Set false for plain-TCP
// self-host Valkey/Redis.
// "tls": false
},
// ── Email (transactional) ───────────────────────────────────────────
// Nodemailer transport — used for password resets, email confirmation, etc.
"email": {
"from": "\"Puter\" <no-reply@puter.com>",
"host": "smtp.example.com",
"port": 587,
"secure": false,
"service": "",
"auth": {
"user": "",
"pass": ""
}
},
// ── Alarms / alerting ───────────────────────────────────────────────
// Where system alarms go. Severity is the routing decision — each
// transport takes everything at or above its own `minSeverity`:
//
// critical — an unhandled server error; pages on-call.
// error — pages as well; prefer critical or warning.
// warning — look at it today; no page.
// info — a record in the chat channel only.
//
// Both transports are off unless enabled, so a self-hosted node just
// logs its alarms to the console.
"pager": {
// Severity for call sites that don't pick one. Default "critical".
"defaultSeverity": "critical",
// Retier or silence an alarm without a deploy. Keys are alarm ids,
// or a prefix ending in `*`; the exact id wins over a pattern, and
// the longest matching pattern wins among patterns. Values are a
// severity or "mute". This is applied last, so it overrides both the
// call site and any known-error rule.
"severityOverrides": {
// "cronMonitor:*": "info",
// "http_500:GET:/some/flapping/route:*": "mute"
},
"pagerduty": {
"enabled": false,
"routingKey": "",
// Lowest severity that reaches PagerDuty. Default "warning",
// which keeps `info` out of the paging system entirely.
"minSeverity": "warning"
},
// Slack incoming webhook — the low-noise destination for everything
// that shouldn't page.
"slack": {
"enabled": false,
"webhookUrl": "",
// Optional; defaults to the channel the webhook was created for.
"channel": "#alerts",
"username": "puter-alarms",
// Severity window posted to Slack. The ceiling defaults to
// "info" when PagerDuty is configured — what pages belongs in
// the pager, not in chat — and to "critical" when Slack is the
// only transport.
"minSeverity": "info",
"maxSeverity": "info",
// Don't repost the same alarm id within this window. The first
// occurrence always posts; the next post that gets through
// reports how many occurrences piled up. 0 disables throttling.
"repeatThrottleMs": 900000
}
},
// ── Rate limiting ───────────────────────────────────────────────────
// `memory` for single-node, `redis` for multi-node (default), `kv` for
// dynamo-backed counters.
"rate_limit": {
"backend": "redis"
},
// ── AI / integration providers ──────────────────────────────────────
// All AI drivers (chat, image, video, TTS, OCR, STT) read from here.
// Provider id == driver-side identifier. Leave empty / omit to disable.
"providers": {
// ─ Chat / completion ─
"claude": { "apiKey": "" },
"openai-completion": { "apiKey": "" },
"azure-openai": {
"apiKey": "",
"apiURL": ""
},
"gemini": { "apiKey": "" },
"groq": { "apiKey": "" },
"deepseek": { "apiKey": "" },
"mistral": { "apiKey": "" },
"xai": { "apiKey": "" },
"moonshot": { "apiKey": "" },
"minimax": {
"apiKey": "",
"apiBaseUrl": "https://api.minimax.io/v1"
},
"openrouter": {
"apiKey": "",
"apiBaseUrl": "https://openrouter.ai/api/v1"
},
"infron": {
"apiKey": "",
"apiBaseUrl": "https://llm.onerouter.pro/v1"
},
// BytePlus ModelArk. One key powers chat (Seed/GLM/DeepSeek), image
// generation (Seedream) and video generation (Seedance); `apiBaseUrl`
// selects the region; see
// https://docs.byteplus.com/en/docs/ModelArk/1330310 for options.
"byteplus": {
"apiKey": "",
"apiBaseUrl": "https://ark.ap-southeast.bytepluses.com/api/v3"
},
"zai": { "apiKey": "" },
"alibaba": { "apiKey": "" },
"together-ai": { "apiKey": "" },
// Local Ollama. `enabled: false` skips the auto-probe at startup
// (otherwise Puter logs ECONNREFUSED on every boot when no Ollama
// is running).
"ollama": {
"enabled": false,
"apiBaseUrl": "http://localhost:11434"
},
// ─ Image generation ─
"openai-image-generation": { "apiKey": "" },
"gemini-image-generation": { "apiKey": "" },
"together-image-generation": { "apiKey": "" },
"cloudflare-image-generation": {
"apiToken": "",
"accountId": "",
"apiBaseUrl": "https://api.cloudflare.com/client/v4"
},
"xai-image-generation": { "apiKey": "" },
// ─ Video generation ─
"openai-video-generation": { "apiKey": "" },
"together-video-generation": { "apiKey": "" },
"gemini-video-generation": { "apiKey": "" },
// ─ Speech / OCR ─
"openai": { "apiKey": "" },
"elevenlabs": {
"apiKey": "",
"apiBaseUrl": "https://api.elevenlabs.io",
"defaultVoiceId": "",
"speechToSpeechModelId": ""
},
"aws-polly": {
"access_key": "",
"secret_key": "",
"region": "us-west-2"
},
"speechify": { "apiKey": "" },
"aws-textract": {
"access_key": "",
"secret_key": "",
"region": "us-west-2"
},
"mistral-ocr": { "apiKey": "" }
},
// ── GUI / static mounts ─────────────────────────────────────────────
"gui_assets_root": "./src/gui",
"gui_profile": "development",
"builtin_apps": {
"dev-center": "./src/dev-center"
},
// Force the bundled GUI even in dev — set true when running from a
// pre-built tree without webpack-dev-server.
"use_bundled_gui": false,
"gui_bundle": "/dist/bundle.min.js",
"gui_css": "/dist/bundle.min.css",
"gui_puterjs_bundle": "https://js.puter.com/v2/",
"gui_params": {
"title": "Puter",
"short_description": "Your personal cloud computer",
"social_media_image": ""
},
// Optional roots for native app bundles and custom puter.js builds.
"native_apps_root": "",
"client_libs_root": "",
"puterjs_root": "./src/puter-js/dist",
// ── Feature flags (whoami extension) ────────────────────────────────
// Flat `{ flag_name: boolean }` bag. Server-only by default — flags are
// only surfaced to the client if their key is on the allowlist in
// `extensions/whoami.ts` (CLIENT_VISIBLE_FEATURE_FLAGS).
"feature_flags": {
"example_flag": false
},
// ── Misc / safety ───────────────────────────────────────────────────
// TLDs / domains rejected at signup (prod only).
"blockedEmailDomains": [],
"support_email": "support@puter.com",
// Worker / subdomain names users can't claim.
"reserved_words": [],
"max_subdomains_per_user": 10,
"server_health": {
"db_liveness_latency_fail_ms": 1500,
"stale_health_loop_fail_ms": 0
},
// ── Extensions ──────────────────────────────────────────────────────
// Directories scanned for extension entrypoints (`*.ts` / subdirs).
"extensions": [
"./extensions"
],
// ── Metering ────────────────────────────────────────────────────────
// When true, every account resolves to an unlimited policy: usage is still
// recorded, but nothing is ever refused for lack of budget. This is the
// setting for a deployment with no way to buy more — without it, accounts
// are held to the free monthly allowance and start getting 402s from the AI
// surfaces, file transfers and KV once they pass it.
"unlimitedMetering": false,
// Whether an account that has spent its whole allowance is refused the
// operations that spend it — file transfers, KV calls. Recording is
// unaffected either way. `workers` extends the same refusal to
// worker-driven calls, which are exempt by default because a deployed
// worker has nowhere to surface a payment prompt. `subscriptions` covers
// the separate question of surfaces reserved for paid plans (the
// vendor-compatible AI endpoints): turn it off on a deployment with no
// paid plans, or every account is refused them.
// "meteringEnforcement": { "enabled": true, "workers": false, "subscriptions": true },
// Fleet-wide spend rate, in micro-cents per minute, past which metering
// raises the `metering:excessiveGlobalUsageRate` alarm. Omit it (the
// default) to leave the check off: the only useful value is a multiple of
// what this deployment's normal traffic costs, so it has to be measured
// rather than guessed, and a stale number here alarms on healthy growth.
// "maxGlobalUsagePerMinute": 200000000
}
Reference in new issue View Git Blame Copy Permalink