Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
HOSTNoHTTP listening host.0.0.0.0
PORTNoHTTP listening port.8099
HEADLESSNoDocker image runs with Xvfb in visible mode. Headless mode triggers verification more often.false
MCP_PATHNoHTTP path for the MCP endpoint./mcp
LOG_LEVELNoLog level. Logs are written to stderr as JSON; cookie and token fields are masked.info
DIAGNOSTICSNoSave a screenshot on error (available via `check_session` as `last_error_screenshot`).true
RETURN_MODENoResult mode: `both`, `image`, or `link`.both
MCP_TRANSPORTNoTransport to use: `http` or `stdio`. The `--stdio` argument also works.http
BROWSER_LOCALENoBrowser locale fingerprint.en-US
MAX_STORAGE_MBNoLibrary storage limit in MB. `0` disables the limit.2048
MCP_AUTH_TOKENNoIf set, requires `Authorization: Bearer` for /mcp. Not needed on Obot because Obot handles authentication.
RETENTION_DAYSNoLibrary cleanup retention in days. `0` disables the limit.14
BROWSER_CHANNELNoUse a specific Chrome or Chromium channel.
CHATGPT_CDP_URLNoConnect to an existing Chrome instead of launching a browser. Remote addresses require ALLOW_REMOTE_CDP=true.
PERSIST_SESSIONNoPersist refreshed cookies under /data. If the env value changes, the new value takes precedence.true
PUBLIC_BASE_URLNoEnables signed /files/<id> download links.
ALLOW_REMOTE_CDPNoAllow remote CDP addresses when connecting via CHATGPT_CDP_URL.
ALLOW_URL_INPUTSNoAllow downloading https image URLs. Private network addresses are always blocked.true
BROWSER_TIMEZONENoBrowser timezone fingerprint.UTC
INLINE_MAX_BYTESNoTotal size limit for images returned inline.4 MiB
MAX_INPUT_IMAGESNoMaximum number of input images.8
MAX_QUEUE_LENGTHNoMaximum number of queued jobs.20
SYNC_WAIT_SECONDSNoHow long a tool call waits for a result before returning a job_id. Keep below 60 seconds for clients that time out tool calls at 60 seconds.45
ALLOWED_INPUT_DIRSNoDirectories allowed for local file inputs. If empty, local files are not read.
BROWSER_USER_AGENTNoBrowser user agent fingerprint.
MAX_IMAGES_PER_JOBNoUpper limit for the `count` parameter.4
CHATGPT_DEFAULT_URLNoURL where new conversations are opened, for example a ChatGPT Project or GPT.https://chatgpt.com/
CHATGPT_PROFILE_DIRNoUse a persistent Chromium profile instead of an ephemeral browser context.
MAX_CONCURRENT_JOBSNoNumber of concurrently running tabs (maximum 4).1
BROWSER_PROXY_SERVERNoBrowser proxy server.
CHATGPT_SESSION_STATENoOutput from `npm run export-session`. Can be base64, plain JSON, a file path, or a cookie-extension JSON array.
CHATGPT_SESSION_TOKENNoValue of the __Secure-next-auth.session-token cookie. Chunked cookies can be given as `part0|part1`; multiple cookies as `name=value; name2=value2`.
MAX_INPUT_IMAGE_BYTESNoMaximum size per input image.20 MiB
BROWSER_PROXY_PASSWORDNoBrowser proxy password.
BROWSER_PROXY_USERNAMENoBrowser proxy username.
CHATGPT_SELECTORS_JSONNoOverride ChatGPT UI selectors when the interface changes. See src/browser/selectors.ts for keys.
BROWSER_EXECUTABLE_PATHNoPath to a specific Chrome or Chromium executable.
GENERATION_TIMEOUT_SECONDSNoMaximum wait time for a single image generation.300

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

Server capabilities have not been inspected yet.

Tools

Functions exposed to the LLM to take actions

NameDescription

No tools

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources