gpt4free — Settings

Account — g4f.space

Sign in to sync conversations across devices via your server's /v1/secret/conversations endpoints (AES-256-GCM encrypted at rest).

Server

Server URL

Defaults to the public https://g4f.space instance. For a local server run python -m g4f --port 1337 and enter http://localhost:1337.

API key (optional)

Sent as Authorization: Bearer … and g4f-api-key.

Browser as CDP provider

Let your g4f server drive this browser (automation tabs via chrome.debugger) instead of launching its own headless Chrome. Requires a local g4f server with G4F_BROWSER_MODE=extension.

Chat defaults

Model

Loaded from /v1/models. Leave empty to let the server pick.

Provider (optional)

System prompt

Temperature (empty = default)

Max tokens (empty = default)

Stream responses token by token

Images

Image model

Images per request

Generate images from the side panel with /img your prompt.

Context & behavior

Page context limit (characters)

Send on Enter (Shift+Enter for newline)

Quick actions

Shown as buttons in the popup. {text} is replaced with the selection/page content.