How do I use unseen-mcp?

1. Click on "Deploy Server". 2. Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state. 3. In the chat, type @ followed by the MCP server name and your instructions, e.g., "@unseen-mcp turn this incident playbook into a cited visual workflow" That's it! The server will respond to your query, and you can continue using it as needed. Here is a step-by-step guide with screenshots.

unseen-mcp

by eyeinthesky6

Overview Schema Source Related Servers Discussions Admin

JavaScript

Remote

Unseen MCP Server (`unseen-mcp`)

The official Model Context Protocol (MCP) server for Unseen (unseendocs.in).

Turn messy operational documents, standard operating procedures (SOPs), incident playbooks, spreadsheets, and business communications into faithful, source-grounded visual workflows directly inside Claude Desktop, Cursor, Zed, Windsurf, or any MCP-compatible agent.

Capabilities

The unseen-mcp server provides 3 primary tools:

compile_workflow: Compiles raw text, SOPs, or multi-page documents into a structured, verifiable VisualSpec workflow graph, accompanied by production-ready SVG and Mermaid syntax.
- Arguments: content (string, required), title (string, optional).
revise_workflow: Applies focused revisions or node additions/modifications to an existing workflow graph while preserving citation integrity.
- Arguments: visualSpec (object, required), instructions (string, required).
get_workflow: Retrieves the full structured representation of an existing visual artifact by its ID.
- Arguments: id (string, required).

Related MCP server: obscuraai-mcp

1-Click Installation (Smithery)

If you use Smithery.ai:

npx -y @smithery/cli install unseen-mcp --client claude

Or for Cursor:

npx -y @smithery/cli install unseen-mcp --client cursor

Claude Desktop Configuration

Add the following to your claude_desktop_config.json:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
Windows: %APPDATA%\Claude\claude_desktop_config.json

{
  "mcpServers": {
    "unseen": {
      "command": "npx",
      "args": ["-y", "unseen-mcp"],
      "env": {
        "OPENROUTER_API_KEY": "sk-or-v1-your-openrouter-key",
        "TYPESAFE_API_KEY": "your-optional-typesafe-key"
      }
    }
  }
}

Cursor IDE Configuration

In Cursor Settings -> Features -> MCP:

Click + Add New MCP Server.
Name: unseen
Type: command
Command: npx -y unseen-mcp
Add environment variables:
- OPENROUTER_API_KEY: Your OpenRouter API key

Environment Variables

Variable	Description	Required?
`OPENROUTER_API_KEY`	OpenRouter API key for LLM-assisted document compilation	Recommended for rich natural language compilation (falls back to deterministic heuristics if absent)
`TYPESAFE_API_KEY`	TypeSafe API key for Jev semantic audits and missing branch recovery	Optional

Running Locally

To run the stdio server directly from the command line:

npx -y unseen-mcp

Or build from source in the unseen repository:

git clone https://github.com/eyeinthesky6/unseen.git
cd unseen
npm install
npm run build:mcp
node dist-mcp/index.cjs

License

MIT © Unseen

Available Tools

9 tools

auth_loginSign In with Email + PasswordInspect

Authenticate an existing Unseen account using email and password (or create one if register: true), returning { token: 'utok_...', user } for use in Authorization: Bearer <token> or authToken.

ParametersJSON Schema

Name	Required	Description
`name`	No	Optional display name when registering.
`email`	Yes	The user's email address.
`password`	Yes	Account password (min 8 characters).
`register`	No	Set to true to register a new Email + Password account if it does not exist yet.

auth_request_otpRequest Agent & Human Login OTP (Email + OTP)Inspect

Step 1 of Agent & Human Login (Email + OTP): Sends a 6-digit Login OTP (UNSEEN_LOGIN_OTP) to the user's email address. This is a passwordless Login OTP (NOT a password reset). Once called, the agent can either read the 6-digit UNSEEN_LOGIN_OTP directly from the user's email (when the user has granted email access) or ask the user in chat for the 6-digit code, then call auth_verify_otp.

ParametersJSON Schema

Name	Required	Description	Default
`email`	Yes	The user's email address to receive the 6-digit Login OTP.

auth_verify_otpVerify Agent & Human Login OTPInspect

Step 2 of Agent & Human Login (Email + OTP): Verifies the 6-digit Login OTP sent to email and returns { token: 'utok_...', user } (30 free credits / 3 visuals per month on new accounts). Pass the returned token as Authorization: Bearer <token> or in the authToken argument of subsequent MCP tool calls.

ParametersJSON Schema

Name	Required	Description	Default
`otp`	Yes	The 6-digit Login OTP read from the user's email (`UNSEEN_LOGIN_OTP`) or supplied by the user.
`email`	Yes	The user's email address.

autofix_workflowAutofix & Re-Render Workflow ($0 Free)A

Idempotent

Inspect

Deterministic connector healing and full visual re-rendering ($0 compute cost, instant, unlimited). Rescues orphan nodes, heals broken or missing decision branches, recomputes deterministic orthogonal layout, and returns fresh spec, svg, mermaid, excalidraw, and studioUrl. Use this tool after making direct structural edits to VisualSpec JSON (nodes / edges), or when a user asks to autofix connectors.

ParametersJSON Schema

Name	Required	Description
`spec`	Yes	The VisualSpec JSON object to deterministically heal and re-render into SVG, Mermaid, and Excalidraw.
`authToken`	No	Optional session token (`utok_...`) if not passed via Authorization header.
`orientation`	No	Optional layout orientation (defaults to portrait).

TDQS

A4/5.0

Behavior4/5

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

Annotations already declare idempotent, non-destructive and closed-world; the description adds real context beyond that: deterministic healing of orphan nodes and broken/missing decision branches, layout recomputation, zero compute cost/unlimited calls, and that a fresh spec is returned rather than mutating in place. It stops short of describing failure behavior on invalid specs or whether caller-supplied layout is discarded.

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?

Two sentences, front-loaded with the core action before the return values and the usage trigger. Slightly diluted by promotional framing ('$0 compute cost, instant, unlimited') that partially duplicates the title and the parenthetical.

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?

With no output schema, the description compensates by enumerating the exact returned artifacts, and for a three-parameter healing tool it covers purpose, triggers and side effects. Nothing an agent needs to invoke it correctly appears missing.

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

Parameters3/5

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

Schema description coverage is 100% across all three parameters, so spec/authToken/orientation are already documented in the schema. The description only loosely restates that spec is the object to heal and implies layout options via 'recomputes deterministic orthogonal layout', adding little parameter-level meaning. Baseline 3 applies when the schema carries the load.

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

Purpose4/5

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

States a specific action set: deterministic connector healing plus full re-rendering, and enumerates the returned artifacts (spec, svg, mermaid, excalidraw, studioUrl). It is distinct from compile/revise/enhance by being a heal-and-rerender pass, but the description never names those siblings explicitly, so differentiation is inferred rather than stated.

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 two concrete triggers: after direct structural edits to VisualSpec nodes/edges, or when a user asks to autofix connectors. That is a clear when-to-use statement. It does not say when *not* to use it or which sibling (compile_workflow, revise_workflow, enhance_workflow) handles the adjacent cases.

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

compile_workflowCompile Workflow DiagramInspect

Turn supplied document evidence or raw operational text into a publication-ready workflow diagram with source references, clean vector SVG, Mermaid, Excalidraw scene, and interactive Studio deep link (studioUrl). HOW TO STRUCTURE TEXT INPUTS: Either pass sources (array of 1..60 verbatim chunks {id, locator, excerpt} where each excerpt is <= 2000 chars with exact numbers/actors/thresholds preserved) OR pass raw text and Unseen will automatically segment it into source chunks. EDITING GUIDANCE: For cosmetic/styling/annotation edits, edit the returned <svg> XML directly. For structural/topological edits (adding/removing nodes or rerouting branches), edit spec (VisualSpec JSON) and call autofix_workflow ($0 deterministic connector healing + fresh SVG/Mermaid/Excalidraw) or revise_workflow.

ParametersJSON Schema

Name	Required	Description
`text`	No	Optional raw document or user text. If provided when `sources` is omitted or empty, Unseen automatically segments `text` into faithful `sources[]` chunks.
`jobId`	No	Optional job ID when resuming a workflow after clarification
`sources`	Yes	Array of faithful source excerpts with locators from the input document (1..60 chunks, <=2000 chars each). Preserve exact thresholds, actors, and IF/ELSE branches.
`userGoal`	Yes	The overarching operational goal or process description (e.g. 'Driver Onboarding SOP' or 'Vendor Risk Assessment')
`authToken`	No	Optional session token (`utok_...`) obtained from `auth_verify_otp` or `auth_login` if not already supplied in the HTTP `Authorization: Bearer` header.
`requestId`	No	Optional client request correlation ID
`hostContext`	No	Optional advisory context observed by the host agent
`priorAnswers`	No	Answers to prior Unseen clarification questions when resuming a compilation
`desiredOutput`	No
`schemaVersion`	Yes	Protocol envelope version (must be '0.1')

detect_workflowsDetect Candidate Workflows in Document

Read-onlyIdempotent

Inspect

Discover distinct operational workflows from multi-section documents or raw text before compiling. Returns candidate workflows (id, title, description, sourceIds) audited by TypeSafe Jev.

ParametersJSON Schema

Name	Required	Description
`text`	No	Optional raw document text if `sources` is not pre-segmented.
`sources`	No	Array of source excerpts `{id, locator, excerpt}` from the document.
`authToken`	No	Optional session token (`utok_...`) if not passed via Authorization header.

enhance_workflowAI Enhance Workflow (TypeSafe Jev Audit & Repair)Inspect

Bundled free semantic verification and repair powered by TypeSafe Jev. Audits the VisualSpec against its cited sources[] for unsupported claims, missing decision branches, and role accuracy, applies targeted repair if actionable issues are found, and returns the enhanced spec, svg, mermaid, excalidraw, studioUrl, and Jev quality metrics.

ParametersJSON Schema

Name	Required	Description
`spec`	Yes	The current VisualSpec JSON object (including its `sources` array) to audit and enhance.
`visualId`	No	The `jobId` or `visualId` returned when the workflow was compiled.
`authToken`	No	Optional session token (`utok_...`) if not passed via Authorization header.

get_workflowGet Compiled Workflow

Read-onlyIdempotent

Inspect

Retrieve an existing VisualSpec workflow by jobId or visualId, returning the spec, SVG layout, and Mermaid diagram.

ParametersJSON Schema

Name	Required	Description	Default
`id`	Yes	The jobId or visualId of the workflow to retrieve
`authToken`	No	Optional session token (`utok_...`) if not passed via Authorization header.

revise_workflowRevise Workflow Diagram

Idempotent

Inspect

Apply targeted delta revisions to an existing VisualSpec workflow diagram (e.g. modify node labels, adjust roles like step vs decision, add/remove edges, or heal connections) while preserving Visual Contract v0.1 validity and source grounding.

ParametersJSON Schema

Name	Required	Description
`spec`	Yes	The existing VisualSpec object to revise
`authToken`	No	Optional session token (`utok_...`) if not passed via Authorization header.
`instruction`	Yes	Clear revision instruction (e.g. 'Change node-2 role to decision and label to Finance Approval?', 'Remove orphan node-4', 'Autofix layout connections')
`additionalSources`	No	Optional additional source evidence excerpts to support the revision

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

9 tool updatesv1.0.0Oct 23, 2026
- First observedauth_login
- First observedauth_request_otp
- First observedauth_verify_otp
- First observedautofix_workflow
- First observedcompile_workflow
- First observeddetect_workflows
- First observedenhance_workflow
- First observedget_workflow
- First observedrevise_workflow

Deploy Server

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

TDQS

Score is being calculated.

Maintenance

ActivityMaintained

ResponsivenessNo issues

Resources

Need Help?

Related Servers

Related MCP Connectors

Wavelength AOAuth
com.wavelengthflow
Turn calls and notes into process flow diagrams in Claude, ChatGPT, VS Code, Cursor, and Teams.
Note Taking Project Management Software Architecture
1061
AgentDrive A
sh.agentdrive
Cross-agent artifact workspace with provenance across Claude Code, Codex, Cursor, LangGraph.
Agent Orchestration Developer Tools
222
agentdocsOAuth
io.github.dancourse
An agent-first office suite Claude & ChatGPT read and write over one MCP URL.
Workplace & Productivity Documentation Access
2
BrainTubeOAuth
io.github.lildaddyo
Your saved videos, articles and notes as one knowledge base for Claude, ChatGPT and any MCP client.
Search Note Taking Knowledge & Memory
3

Related MCP Servers

Agentled MCP Server
Agent Orchestration AI & Machine Learning Knowledge & Memory
Agentled
A
license
A
quality
D
maintenance
AI workflow orchestration platform with 100+ integrations, knowledge graph, and unified credits. Build, manage, and execute workflows from Claude Code, Cursor, or any MCP client.
Updated 8 days ago (2026-10-15 20:59 UTC)
40
359 npm
3
MIT
obscuraai-mcp
Autonomous Agents App Automation Workplace & Productivity
madeinphantom
A
license
B
quality
D
maintenance
MCP server for Claude, ChatGPT, coding agents (cursor, claude code etc), windsurf, etc— generate visual AI automation workflows. Describe a business process, get an interactive flow map you can edit and share.
Updated 6 months ago (2026-04-10 11:39 UTC)
1
25 npm
MIT
Cortex
Knowledge & Memory Search Note Taking
liveevilkin
A
license
Not graded
quality
D
maintenance
Obsidian-backed knowledge graph with semantic search, entity extraction, and cross-session memory. 11 MCP tools. Works with Claude Code, Cursor, Windsurf, and any MCP-compatible editor.
Updated 4 months ago (2026-06-19 11:03 UTC)
54 npm
1
MIT
Circuit
App Automation Autonomous Agents
brandononchain
A
license
Not graded
quality
B
maintenance
Enables building and running visual workflows conversationally inside Claude, orchestrating your existing connector tools step-by-step without storing credentials or owning integrations.
Updated a month ago (2026-09-20 01:23 UTC)
8 npm
MIT

Unseen MCP Server (unseen-mcp)

Capabilities

1-Click Installation (Smithery)

Claude Desktop Configuration

Cursor IDE Configuration

Environment Variables

Running Locally

License

Available Tools

TDQS

Tool Schema Changelog

TDQS

Maintenance

Resources

Related MCP Connectors

Related MCP Servers

Agentled MCP Server

obscuraai-mcp

Cortex

Circuit

Unseen MCP Server (`unseen-mcp`)