chatgpt-webui-mcp
Server Configuration
Describes the environment variables required to run the server.
| Name | Required | Description | Default |
|---|---|---|---|
| CHATGPT_SESSION_TOKEN | No | Session token for chatgpt.com, obtained from cookies. Required if not using CHATGPT_SESSION_TOKEN_FILE. | |
| CHATGPT_WAIT_TIMEOUT_MS | No | Timeout in milliseconds for waiting (default: 7200000). | |
| CHATGPT_BROWSER_BASE_URL | No | Base URL for the browser (default: http://127.0.0.1:9377). | |
| CHATGPT_SESSION_TOKEN_FILE | No | Path to a file containing the session token. Alternative to CHATGPT_SESSION_TOKEN. | |
| CHATGPT_IMAGE_SCREENSHOT_FALLBACK | No | Set to '1' to enable fallback screenshot output. | |
| CHATGPT_IMAGE_SCREENSHOT_MAX_BYTES | No | Maximum bytes for fallback screenshot (default: 2097152). |
Instructions
Guidance the server publishes about itself, which clients place ahead of the tool catalog so the model reads it before choosing anything.
This server publishes no instructions, or was last inspected before Glama recorded them.
Capabilities
Features and capabilities supported by this server
Protocol revision2025-11-25
| Capability | Details |
|---|---|
| tools | {
"listChanged": true
} |
Tools
Functions exposed to the LLM to take actions
| Name | Description |
|---|---|
| chatgpt_webui_sessionA | Validate ChatGPT session token and return session details. |
| chatgpt_webui_modelsA | List available ChatGPT WebUI models for the current account. |
| chatgpt_webui_askC | Send a prompt to ChatGPT WebUI using session token auth and return assistant text. |
| chatgpt_webui_ask_async_startA | Start a background ChatGPT ask job and return immediately with a job id. Use this for long-running tasks like Deep Research and Pro runs. |
| chatgpt_webui_ask_async_statusC | Get status for a background ask job. |
| chatgpt_webui_ask_async_resultB | Get result for a background ask job. Optionally wait for completion by providing wait_timeout_ms. |
| chatgpt_webui_promptB | Unified prompt entrypoint. mode=auto uses background runs for long tasks (Deep Research, Pro/Thinking, image generation) and direct wait for short tasks. |
| chatgpt_webui_runB | Unified run checker. Use run_id from chatgpt_webui_prompt background mode (or job_id from legacy async tools). |
| chatgpt_webui_commandB | Natural-language command wrapper. Converts phrases like 'with chatgpt webui on gpt 5.2 pro extended thinking: ...' into a chatgpt_webui_prompt call. |
Prompts
Interactive templates invoked by user choice
| Name | Description |
|---|---|
No prompts | |
Resources
Contextual data attached and managed by the client
| Name | Description |
|---|---|
No resources | |
TDQS
Scored across 9 tools
There is significant functional overlap: chatgpt_webui_ask, chatgpt_webui_prompt, chatgpt_webui_command, and chatgpt_webui_ask_async_start all initiate ChatGPT prompts. The distinction between direct ask, unified prompt, and natural-language wrapper is not immediately clear, and the run checker overlaps with async status/result tools.
Tools consistently share the chatgpt_webui_ prefix and use snake_case, but the action style is mixed: ask/start are verbs, while session/models/command/prompt/run are nouns. The async subfamily is internally consistent, but the overall naming pattern is not uniform.
Nine tools is within a reasonable range for a ChatGPT WebUI integration. However, several tools appear redundant due to the coexistence of legacy async functions, a unified prompt entrypoint, and a command wrapper, making the set slightly larger than needed.
The core workflow of sending prompts and retrieving results is covered, including background execution for long-running tasks. Notable gaps include no explicit cancellation/abort for background jobs and no conversation-history or context management, which limits full lifecycle coverage.