GitHubFork/grok2api

Public

WeChat Login

Code Issues Pull requests Events Packages Insights

main

自动同步 GitHub 更新

.github
app
docs
scripts
.cnb.yml
.dockerignore
.env.example
.gitignore
.python-version
Dockerfile
LICENSE
README.md
config.defaults.toml
docker-compose.yml
pyproject.toml
uv.lock

NOTE

This project is for learning and research only. Please comply with Grok's terms of use and all applicable local laws and regulations. Do not use it for illegal purposes. If you fork the project or open a PR, please keep the original author and frontend attribution.

Grok2API is a FastAPI-based Grok gateway that exposes Grok Web capabilities through OpenAI-compatible APIs. Core features:

OpenAI-compatible endpoints: /v1/models, /v1/chat/completions, /v1/responses, /v1/images/generations, /v1/images/edits, /v1/videos, /v1/videos/{video_id}, /v1/videos/{video_id}/content
Anthropic-compatible endpoint: /v1/messages
Streaming and non-streaming chat, explicit reasoning output, function-tool structure passthrough, and unified token / usage accounting
Multi-account pools, tier-aware selection, failure feedback, quota synchronization, and automatic maintenance
Local image/video caching and locally proxied media URLs
Text-to-image, image editing, text-to-video, and image-to-video support
Built-in Admin dashboard, Web Chat, Masonry image generation, and ChatKit voice page

Service Architecture

Quick Start

Local Deployment

git clone https://github.com/chenyme/grok2api
cd grok2api
cp .env.example .env
uv sync
uv run granian --interface asgi --host 0.0.0.0 --port 8000 --workers 1 app.main:app

Docker Compose

git clone https://github.com/chenyme/grok2api
cd grok2api
cp .env.example .env
docker compose up -d

Vercel

Render

First Launch

Change app.app_key
Set app.api_key
Set app.app_url otherwise image and video URLs may return 403 Forbidden

WebUI

Routes

Page	Path
Admin login	`/admin/login`
Account management	`/admin/account`
Config management	`/admin/config`
Cache management	`/admin/cache`
WebUI login	`/webui/login`
Web Chat	`/webui/chat`
Masonry	`/webui/masonry`
ChatKit	`/webui/chatkit`

Authentication Rules

Scope	Config	Rule
`/v1/*`	`app.api_key`	No extra authentication when empty
`/admin/*`	`app.app_key`	Default value: `grok2api`
`/webui/*`	`app.webui_enabled`, `app.webui_key`	Disabled by default; if `webui_key` is empty, no extra verification is required

Configuration

Configuration Layers

Location	Purpose	Effective Time
`.env`	Pre-start configuration	At service startup
`${DATA_DIR}/config.toml`	Runtime configuration	Effective immediately after save
`config.defaults.toml`	Default template	On first initialization

Environment Variables

Variable	Description	Default
`TZ`	Time zone	`Asia/Shanghai`
`LOG_LEVEL`	Log level	`INFO`
`LOG_FILE_ENABLED`	Write local log files	`true`
`ACCOUNT_SYNC_INTERVAL`	Account directory incremental sync interval in seconds	`30`
`ACCOUNT_SYNC_ACTIVE_INTERVAL`	Active sync interval after account-directory changes are detected, in seconds	`3`
`SERVER_HOST`	Service bind address	`0.0.0.0`
`SERVER_PORT`	Service port	`8000`
`SERVER_WORKERS`	Granian worker count	`1`
`HOST_PORT`	Docker Compose published host port	`8000`
`DATA_DIR`	Local data root for accounts, locally cached media files, and cache indexes	`./data`
`LOG_DIR`	Local log directory	`./logs`
`ACCOUNT_STORAGE`	Account storage backend	`local`
`ACCOUNT_LOCAL_PATH`	SQLite path for `local` account storage	`${DATA_DIR}/accounts.db`
`ACCOUNT_REDIS_URL`	Redis DSN for `redis` mode	`""`
`ACCOUNT_MYSQL_URL`	SQLAlchemy DSN for `mysql` mode	`""`
`ACCOUNT_POSTGRESQL_URL`	SQLAlchemy DSN for `postgresql` mode	`""`
`ACCOUNT_SQL_POOL_SIZE`	Core connection pool size for SQL backends	`5`
`ACCOUNT_SQL_MAX_OVERFLOW`	Maximum overflow connections above pool size	`10`
`ACCOUNT_SQL_POOL_TIMEOUT`	Seconds to wait for a free connection from the pool	`30`
`ACCOUNT_SQL_POOL_RECYCLE`	Max connection lifetime in seconds before reconnect	`1800`
`CONFIG_LOCAL_PATH`	Runtime config file path for `local` config storage	`${DATA_DIR}/config.toml`

Runtime config can also be overridden with GROK_-prefixed environment variables. For example, GROK_APP_API_KEY overrides app.api_key, and GROK_FEATURES_STREAM overrides features.stream.

System Configuration Groups

Group	Key Items
`app`	`app_key`, `app_url`, `api_key`, `webui_enabled`, `webui_key`
`logging`	`file_level`, `max_files`
`features`	`temporary`, `memory`, `stream`, `thinking`, `auto_chat_mode_fallback`, `thinking_summary`, `dynamic_statsig`, `enable_nsfw`, `show_search_sources`, `custom_instruction`, `image_format`, `imagine_public_image_proxy`, `video_format`
`proxy.egress`	`mode`, `proxy_url`, `proxy_pool`, `resource_proxy_url`, `resource_proxy_pool`, `skip_ssl_verify`
`proxy.clearance`	`mode`, `cf_cookies`, `user_agent`, `browser`, `flaresolverr_url`, `timeout_sec`, `refresh_interval`
`retry`	`reset_session_status_codes`, `max_retries`, `on_codes`
`account.refresh`	`basic_interval_sec`, `super_interval_sec`, `heavy_interval_sec`, `usage_concurrency`, `on_demand_min_interval_sec`
`cache.local`	`image_max_mb`, `video_max_mb`
`chat`	`timeout`
`image`	`timeout`, `stream_timeout`
`video`	`timeout`
`voice`	`timeout`
`asset`	`upload_timeout`, `download_timeout`, `list_timeout`, `delete_timeout`
`nsfw`	`timeout`
`batch`	`nsfw_concurrency`, `refresh_concurrency`, `asset_upload_concurrency`, `asset_list_concurrency`, `asset_delete_concurrency`

Image and Video Formats

Config	Allowed Values
`features.image_format`	`grok_url`, `local_url`, `grok_md`, `local_md`, `base64`
`features.imagine_public_image_proxy`	`true`, `false`
`features.video_format`	`grok_url`, `local_url`, `grok_html`, `local_html`

Supported Models

You can use GET /v1/models to retrieve the currently supported model list.

Chat

Model	mode	tier
`grok-4.20-0309-non-reasoning`	`fast`	`basic`
`grok-4.20-0309`	`auto`	`super`
`grok-4.20-0309-reasoning`	`expert`	`super`
`grok-4.20-0309-non-reasoning-super`	`fast`	`super`
`grok-4.20-0309-super`	`auto`	`super`
`grok-4.20-0309-reasoning-super`	`expert`	`super`
`grok-4.20-0309-non-reasoning-heavy`	`fast`	`heavy`
`grok-4.20-0309-heavy`	`auto`	`heavy`
`grok-4.20-0309-reasoning-heavy`	`expert`	`heavy`
`grok-4.20-multi-agent-0309`	`heavy`	`heavy`
`grok-4.20-fast`	`fast`	`basic`, prefers higher-tier pools
`grok-4.20-auto`	`auto`	`super`, prefers higher-tier pools
`grok-4.20-expert`	`expert`	`super`, prefers higher-tier pools
`grok-4.20-heavy`	`heavy`	`heavy`
`grok-4.3-beta`	`grok-420-computer-use-sa`	`super`

Image

Model	mode	tier
`grok-imagine-image-lite`	`fast`	`basic`
`grok-imagine-image`	`auto`	`super`
`grok-imagine-image-pro`	`auto`	`super`

Image Edit

Model	mode	tier
`grok-imagine-image-edit`	`auto`	`super`

Video

Model	mode	tier
`grok-imagine-video`	`auto`	`super`

API Overview

Endpoint	Auth Required	Description
`GET /v1/models`	Yes	List currently enabled models
`GET /v1/models/{model_id}`	Yes	Retrieve one model
`POST /v1/chat/completions`	Yes	Unified entry point for chat, image, and video
`POST /v1/responses`	Yes	OpenAI Responses API compatible subset
`POST /v1/messages`	Yes	Anthropic Messages API compatible endpoint
`POST /v1/images/generations`	Yes	Standalone image generation endpoint
`POST /v1/images/edits`	Yes	Standalone image editing endpoint
`POST /v1/videos`	Yes	Asynchronous video job creation
`GET /v1/videos/{video_id}`	Yes	Retrieve a video job
`GET /v1/videos/{video_id}/content`	Yes	Fetch the final video file
`GET /v1/files/video?id=...`	No	Fetch a locally cached video
`GET /v1/files/image?id=...`	No	Fetch a locally cached image

API Examples

The examples below use http://localhost:8000.

GET /v1/models

curl http://localhost:8000/v1/models \
  -H "Authorization: Bearer $GROK2API_API_KEY"

Field Notes

Field	Location	Description
`Authorization`	Header	Required when `app.api_key` is non-empty. Use `Bearer <api_key>`

POST /v1/chat/completions

Chat:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-4.20-auto",
    "stream": true,
    "reasoning_effort": "high",
    "messages": [
      {"role":"user","content":"Hello"}
    ]
  }'

Image:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-imagine-image",
    "stream": true,
    "messages": [
      {"role":"user","content":"A cat floating in space"}
    ],
    "image_config": {
      "n": 2,
      "size": "1024x1024",
      "response_format": "url"
    }
  }'

Video:

curl http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-imagine-video",
    "stream": true,
    "messages": [
      {"role":"user","content":"A neon rainy street at night, cinematic slow tracking shot"}
    ],
    "video_config": {
      "seconds": 10,
      "size": "1792x1024",
      "resolution_name": "720p",
      "preset": "normal"
    }
  }'

Field Notes

Field	Description
`messages`	Supports text and multimodal content blocks
`stream`	Whether to stream output; falls back to `features.stream` when omitted
`reasoning_effort`	`none`, `minimal`, `low`, `medium`, `high`, `xhigh`; `none` disables reasoning output
`temperature` / `top_p`	Sampling parameters, default `0.8` / `0.95`
`tools`	OpenAI function tools structure
`tool_choice`	`auto`, `required`, or a specific function tool
`image_config`	Image model parameters
\|_ `n`	`1-4` for `lite`, `1-10` for other image models, `1-2` for edit
\|_ `size`	`1280x720`, `720x1280`, `1792x1024`, `1024x1792`, `1024x1024`
\|_ `response_format`	`url`, `b64_json`
`video_config`	Video model parameters
\|_ `seconds`	`6`, `10`, `12`, `16`, `20`
\|_ `size`	`720x1280`, `1280x720`, `1024x1024`, `1024x1792`, `1792x1024`
\|_ `resolution_name`	`480p`, `720p`
\|_ `preset`	`fun`, `normal`, `spicy`, `custom`

POST /v1/responses

curl http://localhost:8000/v1/responses \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-4.20-auto",
    "input": "Explain quantum tunneling",
    "instructions": "Keep the answer concise.",
    "stream": true,
    "reasoning": {
      "effort": "high"
    }
  }'

Field Notes

Field	Description
`model`	Model ID. It must be an enabled model
`input`	User input; supports a string or a Responses API-style message array
`instructions`	Optional system instructions injected as a system message
`stream`	Whether to stream output; falls back to `features.stream` when omitted
`reasoning`	Optional reasoning configuration
\|_ `effort`	`none` disables reasoning output; other values enable it
`temperature` / `top_p`	Sampling parameters, default `0.8` / `0.95`
`tools` / `tool_choice`	Function tools are supported; flat Responses API tool definitions are normalized automatically

POST /v1/messages

curl http://localhost:8000/v1/messages \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-4.20-auto",
    "stream": true,
    "thinking": {
      "type": "enabled",
      "budget_tokens": 1024
    },
    "messages": [
      {
        "role": "user",
        "content": "Explain quantum tunneling in three sentences"
      }
    ]
  }'

Field Notes

Field	Description
`model`	Model ID. It must be an enabled model
`messages`	Anthropic Messages-format messages; supports text, image, document, and tool-result blocks
`system`	Optional system prompt; accepts a string or an array of text blocks
`stream`	Whether to stream output; falls back to `features.stream` when omitted
`thinking`	Optional reasoning configuration
\|_ `type`	`disabled` disables reasoning output; other configs enable it
`max_tokens`	Accepted but currently ignored because Grok upstream does not expose this parameter
`tools` / `tool_choice`	Anthropic tool definitions are supported and converted to internal function tools

POST /v1/images/generations

curl http://localhost:8000/v1/images/generations \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -d '{
    "model": "grok-imagine-image",
    "prompt": "A cat floating in space",
    "n": 1,
    "size": "1792x1024",
    "response_format": "url"
  }'

Field Notes

Field	Description
`model`	Image model: `grok-imagine-image-lite`, `grok-imagine-image`, or `grok-imagine-image-pro`
`prompt`	Image generation prompt
`n`	Number of images; `1-4` for `lite`, `1-10` for other image models
`size`	Supports `1280x720`, `720x1280`, `1792x1024`, `1024x1792`, `1024x1024`
`response_format`	`url` or `b64_json`

POST /v1/images/edits

curl http://localhost:8000/v1/images/edits \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -F "model=grok-imagine-image-edit" \
  -F "prompt=Make this image sharper" \
  -F "image[]=@/path/to/image.png" \
  -F "n=1" \
  -F "size=1024x1024" \
  -F "response_format=url"

Field Notes

Field	Description
`model`	Image-edit model, currently `grok-imagine-image-edit`
`prompt`	Edit instruction
`image[]`	Reference image multipart file field; up to 5 images are used
`n`	Number of outputs, range `1-2`
`size`	Currently only `1024x1024` is supported
`response_format`	`url` or `b64_json`
`mask`	Not supported yet; passing it returns a validation error

POST /v1/videos

curl http://localhost:8000/v1/videos \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -F "model=grok-imagine-video" \
  -F "prompt=A neon rainy street at night, cinematic slow tracking shot" \
  -F "seconds=10" \
  -F "size=1792x1024" \
  -F "resolution_name=720p" \
  -F "preset=normal" \
  -F "input_reference[]=@/path/to/reference.png"

curl http://localhost:8000/v1/videos/<video_id> \
  -H "Authorization: Bearer $GROK2API_API_KEY"

curl -L http://localhost:8000/v1/videos/<video_id>/content \
  -H "Authorization: Bearer $GROK2API_API_KEY" \
  -o result.mp4

Field Notes

Field	Description
`model`	Video model, currently `grok-imagine-video`
`prompt`	Video generation prompt
`seconds`	Video length: `6`, `10`, `12`, `16`, `20`
`size`	Supports `720x1280`, `1280x720`, `1024x1024`, `1024x1792`, `1792x1024`
`resolution_name`	`480p` or `720p`
`preset`	`fun`, `normal`, `spicy`, `custom`
`input_reference[]`	Optional image-to-video reference multipart file field; at most the first 7 images are used
`video_id`	Video job ID returned by `POST /v1/videos`; used to retrieve the job or download the final video