Skip to main content
Glama

Server Configuration

Describes the environment variables required to run the server.

NameRequiredDescriptionDefault
JEV_PROVIDERNoForce Jev provider: 'typesafe' or 'openrouter' instead of auto-detection.auto
GEMINI_API_KEYNoGoogle Gemini API key (starting with AIza) for the typing model.
OPENAI_API_KEYNoOpenAI API key (starting with sk-) for the typing model.
TYPESAFE_API_KEYNoTypeSafe API key. Default provider when set. Obtainable from console.typesafe.ai/settings/keys.
ANTHROPIC_API_KEYNoAnthropic API key (starting with sk-ant-) for the typing model.
JEV_BROWSER_MODELNoPin a Jev version, or 'typesafe/jev-1.13' on OpenRouter.jev-latest
JEV_BROWSER_HEADEDNoSet to '1' to watch the browser.
OPENROUTER_API_KEYNoOpenRouter API key. Powers both Jev judgments and the typing model if no other typing provider is configured. One key runs everything.
JEV_BROWSER_TYPE_MODELNoOverride the typing model, e.g. 'openrouter:anthropic/claude-haiku-4.5'.
JEV_BROWSER_TYPE_API_KEYNoAPI key for the typing endpoint if required.
JEV_BROWSER_TYPE_BASE_URLNoBase URL for any OpenAI-compatible endpoint for the typing model (e.g. Ollama, LM Studio).
JEV_BROWSER_TYPE_PROVIDERNoForce typing provider: 'openai', 'openrouter', 'anthropic', or 'google'.
GOOGLE_GENERATIVE_AI_API_KEYNoAlternate Google API key for the typing model.
JEV_BROWSER_SKIP_BROWSER_DOWNLOADNoSet to '1' to skip the Chromium postinstall download.

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

CapabilityDetails
tools
{
  "listChanged": false
}

Tools

Functions exposed to the LLM to take actions

NameDescription
jev_navigateA

Give a task and a start URL; a Jev-driven agent navigates a real headless browser until the goal is met, the stuck gate fires, or a budget (steps/seconds) is exhausted. Returns the final page in a chosen format (text, markdown, html, or an aria snapshot), the full step trace with confidences, console/page/network errors captured along the way, token usage with estimated cost, and a final screenshot. The result also reports typing degradation explicitly (degraded, warnings with codes, typing_provider, typing_model), so a failed typing generator is visible instead of silently typing keyword soup. For logins: with JEV_BROWSER_PASSWORD_ORIGIN set in this server's environment, password_file or password_env fills native password fields on that origin only, without the value ever entering model context, traces, or screenshots; never put the password value itself in any argument or in the task. To start already logged in, seed a session cookie instead via cookie_file or cookie_env (same reference-based delivery and redaction).

Prompts

Interactive templates invoked by user choice

NameDescription

No prompts

Resources

Contextual data attached and managed by the client

NameDescription

No resources

TDQS

A4.6/5.0

Scored across 1 tool

Disambiguation5/5

Only one tool exists, so there is no risk of misselection or overlapping purposes; its role as the sole browser-agent entry point is unmistakable.

Naming Consistency5/5

The single tool name jev_navigate follows a clear prefix+verb pattern; with no other tools, there is no inconsistency to detect.

Tool Count4/5

A single tool for a browser-agent server is lean but arguably appropriate because it encapsulates navigation, tracing, auth, and reporting in one high-level operation; however it falls below the typical 3-15 range.

Completeness4/5

The tool covers the full navigate-to-goal lifecycle including results, traces, errors, typing degradation, and auth via password/cookie; only fringe browser operations such as direct upload/download or multi-tab control are unaddressed.

Maintenance

ActivityMaintained
ResponsivenessWithin a week