Sakrylle API

English

简体中文

English

简体中文

Appearance

LobeChat ​

Introduction ​

LobeChat is an open-source multi-model AI chat web UI maintained by LobeHub. It supports Web, desktop (Tauri), and self-hosted Docker deployments, and natively offers "custom endpoint" configuration for OpenAI / Anthropic / Google and other providers. Sakrylle API is OpenAI-protocol-compatible and can be plugged in as one of the custom OpenAI providers.

Prerequisites ​

  1. A working LobeChat instance: official online, self-hosted Docker, or a local dev build.
  2. An account at https://ai1.sakrylle.com, with a sk-xxxxxxxxxxxxxxxx key created per API Keys.
  3. The model IDs you want to enable (e.g. gpt-5.6-sol, claude-haiku-4-5-20251001).

Configuration ​

LobeChat UI strings: "Settings", "AI Providers", "OpenAI", "API Key", "API Proxy Address".

  1. Open LobeChat and click the avatar at the top right or Settings at the bottom of the left sidebar.

  2. In the left section, choose AI Providers.

  3. Find OpenAI in the list and click to expand it.

  4. In the OpenAI panel, fill in per this table:

    FieldValue
    API Keysk-xxxxxxxxxxxxxxxx
    API Proxy Address (Base URL)https://api.sakrylle.com/v1
    Use Client ModeRecommended on — direct browser-to-API calls

  5. In the Model List section of the same panel, click Fetch Model List and wait for the request to finish. If the model you want doesn't show up automatically, click + Add Custom Model and enter the full model ID (e.g. gpt-5.6-sol, claude-haiku-4-5-20251001), then save.

  6. Toggle Enabled at the top of the panel to make sure the OpenAI provider is on.

  7. Return to the chat view, tap the model picker at the top of the conversation, and switch the default model to a Sakrylle model you just added.

Tip: if you have multiple providers enabled in LobeChat and want clearer naming, change the Display Name in the OpenAI panel to Sakrylle so it's easy to recognize in the model picker.

Env preset ​

If you self-host LobeChat with Docker / Vercel, you can inject environment variables at deploy time so users don't have to fill in the form repeatedly:

bash
docker run -d -p 3210:3210 \
  -e OPENAI_API_KEY=sk-xxxxxxxxxxxxxxxx \
  -e OPENAI_PROXY_URL=https://api.sakrylle.com/v1 \
  -e OPENAI_MODEL_LIST=+gpt-5.6-sol,+claude-haiku-4-5-20251001 \
  lobehub/lobe-chat

OPENAI_MODEL_LIST uses +modelID to add models; for the full syntax see the official model list configuration docs.

FAQ ​

  • Empty model list: check whether https://api.sakrylle.com/v1/models returns data. An empty list typically means the key has not been assigned a model group.
  • 401 / Invalid API Key: key is wrong or disabled — go back to the Console to regenerate.
  • Streaming output stutters or shows nothing: try toggling "Use Client Mode" off and on. In self-hosted setups, ensure the reverse proxy (Nginx / Cloudflare) does not buffer SSE — set proxy_buffering off.

Official Documentation ​