Local MCP

Web Navigate

web_navigate

Destructive

Navigates a web session to a URL (using its saved login if any) and returns the resulting URL + page title. Opens the session if it doesn't exist. CHECK loaded before reading: when it is false the page did not load and message says why — reading the session then describes a blank document, not the site. Read the page with web_read. When the call comes through the cloud connector, an address of this Mac or of its local network (localhost, 127.0.0.1, 192.168.x.x…) is refused with relay_local_address, for every web_* tool; an AI app running on the Mac itself is not limited.

Input Schema

Name	Required	Description
`url`	Yes	URL to open (https).
`session`	No	Session name (default 'default').
`timeout_seconds`	No	Max seconds to wait for load (default 25).

Output Schema

Name	Required	Description
`ok`	No	True only when the requested page actually loaded.
`url`	No	The URL the browser settled on (redirects followed). If it isn't the page you asked for, the navigation didn't deliver it.
`error`	No	Present only on failure: nav_failed (the browser rejected the load), nav_timeout, or nav_blank (it finished on a blank document).
`title`	No	The resulting page title.
`loaded`	No	True only when the browser ended on a real web page. False means nothing was loaded — do NOT read the session and treat the result as page content.
`message`	No	Present only on failure: why the page didn't load, and what to do about it.
`settle_rejected_by`	No	Present only with nav_timeout: which check ruled out 'the page arrived and WebKit just did not say so' - navigation_started (another navigation began and hung, so the visible document is the old one), other_url (it settled on a different page), or not_complete (the document never finished parsing).

Schema Changelog

Changes observed during successful MCP inspections.

Changed1 schema field changedSep 8, 2026

addedOutput schema / properties / settle_rejected_by

Added value: +{
+  "description": "Present only with nav_timeout: which check ruled out 'the page arrived and WebKit just did not say so' - navigation_started (another navigation began and hung, so the visible document is the old one), other_url (it settled on a different page), or not_complete (the document never finished parsing).",
+  "type": "string"
+}

Changed5 schema fields changedAug 20, 2026

addedOutput schema / properties / error

Added value: +{
+  "description": "Present only on failure: nav_failed (the browser rejected the load), nav_timeout, or nav_blank (it finished on a blank document).",
+  "type": "string"
+}

addedOutput schema / properties / loaded / description

Added value: +"True only when the browser ended on a real web page. False means nothing was loaded — do NOT read the session and treat the result as page content."

addedOutput schema / properties / message

Added value: +{
+  "description": "Present only on failure: why the page didn't load, and what to do about it.",
+  "type": "string"
+}

changedOutput schema / properties / ok / description

Previous value: -"True when the page finished loading."New value: +"True only when the requested page actually loaded."

changedOutput schema / properties / url / description

Previous value: -"The URL after navigation (redirects followed)."New value: +"The URL the browser settled on (redirects followed). If it isn't the page you asked for, the navigation didn't deliver it."

Changed1 schema field changedJul 11, 2026

changedOutput schema / (root)

Previous value: -nullNew value: +{
+  "properties": {
+    "loaded": {
+      "type": "boolean"
+    },
+    "ok": {
+      "description": "True when the page finished loading.",
+      "type": "boolean"
+    },
+    "title": {
+      "description": "The resulting page title.",
+      "type": "string"
+    },
+    "url": {
+      "description": "The URL after navigation (redirects followed).",
+      "type": "string"
+    }
+  },
+  "type": "object"
+}

AddedJul 11, 2026

TDQS

A4.6/5.0

Behavior5/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Goes well beyond the annotations by disclosing that the session is auto-created when absent, that saved logins are reused, that a failed load is signaled via `loaded=false` plus a `message`, and that the cloud connector refuses local/private addresses with `relay_local_address` across all web_* tools. The last point is an environment-level constraint that cannot be inferred from any structured field and would otherwise cause confusing failures.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loads the action and return values, then layers the failure-check and network caveats in descending priority — every sentence carries an operational fact. Slightly dense and mixes three unrelated topics (return values, load failure, relay policy) in one block, but nothing is padding.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness5/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a state-changing, open-world navigation tool with an output schema, the description covers what an agent needs: what it mutates (navigates an existing or new session), how failure is reported, and which addresses are blocked through the cloud relay. Return-value detail is backed by the output schema, so no critical gap remains.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters4/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100%, so baseline is 3, but the description adds real meaning for `session`: it persists a login and is created on demand if it doesn't exist. Only `timeout_seconds` gets no elaboration beyond the schema.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource ('Navigates a web session to a URL'), plus the side effect on returns (resulting URL + page title) and the session-creation behavior. It explicitly points to web_read for the reading step, so an agent can separate it from the rest of the web_* family without opening schemas.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines4/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

Gives a clear operational rule — check `loaded` before reading, and use web_read to read the page — which routes the agent to the correct sibling for the follow-up step. It does not cover the other plausible alternatives (web_login for auth, web_session_list, web_click/web_show), so the guidance is clear but not exhaustive.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

All Local MCP tools Inspect

Try in Browser

One endpoint for every MCP toolRayrun runs every MCP server behind one URL and holds the credentials. Each call is decided per client, tool, and argument, then recorded.ray.runAd

Glama MCP Gateway

Add one secure layer between your agents and this server.