LN
LaunchNests Ship
Universal Model Context Protocol (MCP) Standard

Works with the Agent
You Already Use

Zero proprietary lock-in. Connect Claude Code, Cursor, Windsurf, Codex, or Antigravity in 30 seconds via open stdio MCP. Your agent queries live playbooks, operates your local browser, and verifies submissions without leaking secrets.

npx @repo/mcp@latest
Live Interactive Agent Playground

Select Your AI Coding Environment

Select an agent to see real configuration and simulated stdio telemetry:
Claude Code — ship-launchnests stdio bridge
# Simulated execution flow for Claude Code with LaunchNests Ship MCP:
$ claude 'launch my product with ship-launchnests'
● Initializing stdio transport: ship-launchnests MCP server
→ callTool('get_directories', { category: 'developer-tools', minDA: 60 })
✓ Discovered 18 eligible targets: Uneed (DA 58), DevHunt (DA 62), SaaSHub (DA 73)
→ callTool('claim_submission', { platformId: 'devhunt' }) [Lock acquired: 1800s]
● Navigating local browser to devhunt.org/submit via Chrome DevTools...
⚡ Human handoff requested: Cloudflare Turnstile detected. Please approve in browser.
✓ User verified CAPTCHA in local window. Resuming automated form submission...
→ callTool('report_submission', { status: 'submitted', proofUrl: 'https://devhunt.org/tool/my-app' })
✓ HTTP 200 Verified. State transitioned: verified. Audit log recorded.
agent listening on stdio transport...
When a CAPTCHA or OAuth prompt is encountered, Claude Code displays a clean prompt in your shell. Solve it in your open Chrome window, and the agent auto-resumes.
Protocol: MCP v2024-11-05Ready
Zero Credential Transmission Architecture

Programmatic Tools Your Agent Discovers

Instead of forcing an LLM to guess form fields or write blind scripts, LaunchNests provides 7 strictly typed tools directly inside your agent's runtime.

The Verified Submission Pipeline
Zero Credential Risk

How Stdio Protocol Coordinates With Your Local Chrome

Your coding agent never sends passwords, session cookies, or OTP tokens to LaunchNests. Cloud serves evergreen schemas; execution and authentication stay 100% on your workstation.

1Query

Agent queries get_directories for high-DA targets.

2Lock Target

claim_submission prevents concurrent agent overlap.

3Safe Handoff

CAPTCHAs or 2FA pause execution via request_human.

4Verify

State transitions only upon verified live HTTP 200 listing proof.

Passwords and cookies stay exclusively in your local Chrome user profile.
Verified Local
Bit-Rot Protection

Evergreen Cloud Sync vs Static Markdown

Competitors sell you a private git clone of 300 markdown files. When a directory redesigns its form, their static files fail. LaunchNests updates playbooks in cloud real-time.

LaunchNests Dynamic MCP

Pulls real-time constraints via API. Zero local bit-rot.

Static Git Downloads

Static files break when websites update input IDs or tag pickers.

See detailed comparison breakdown

The 7 Verified MCP Tools Reference

These tools are registered automatically into your agent's context window.

@repo/mcp stdio
get_directories
Discovery

Discovers verified launch targets ranked deterministically by relevance to your product's category, pricing, and domain authority.

params: { category?: string, minDA?: number, isFreeOnly?: boolean }
→ PlatformItem[] with active playbook versions
get_platform
Metadata

Retrieves complete metadata for a specific platform: character limits, asset ratios, backlink type (dofollow/nofollow), and monthly traffic.

params: { platformId: string }
→ PlatformMetadata + validation constraints
get_playbook
Execution

Returns the exact step-by-step navigation instructions, DOM form field IDs, and submission rules for the target directory.

params: { platformId: string }
→ PlaybookSpec with sequential agent steps
claim_submission
Concurrency

Acquires an atomic distributed lock on the directory submission to prevent parallel subagents from submitting duplicates.

params: { platformId: string, ttlSeconds?: number }
→ { success: boolean, lockId: string, expiresAt: Date }
report_submission
Verification

Records submission evidence (live URL, submission timestamp, screenshots) and triggers the 12-state verification lifecycle.

params: { platformId: string, status: string, proofUrl?: string }
→ SubmissionRecord with HTTP 200 verification status
request_human
Security

Gracefully pauses browser automation to let the human founder handle CAPTCHAs, 2FA codes, or payment approval in their open browser.

params: { platformId: string, reason: string, resumeUrl?: string }
→ HumanHandoffReceipt (waits for approval signal)
get_submissions
Analytics

Queries your product's live distribution audit trail, tracking status across discovered, queued, submitted, and published states.

params: { productId?: string, status?: string }
→ SubmissionStatusSummary with live listing URLs

Protocol Compatibility Matrix

Every major AI coding environment verified against our test harness.

Coding EnvironmentTransport ModeBrowser AutomationHuman HandoffSetup TimeStatus
Claude Codestdio pipeChrome DevTools MCP / LocalNative CLI prompt30 seconds
Verified
Cursor & Composer.cursor/mcp.jsonIntegrated Browser / PuppeteerInline Composer dialog45 seconds
Verified
Windsurf (Cascade)windsurf/mcp_configLocal Browser BridgeSidebar Confirmation45 seconds
Verified
GitHub Codex / Copilotstdio pipe / agent CLIPlaywright / DevToolsTerminal pause1 minute
Verified
Google AntigravityDynamic MCP ServerBrowser Subagent sessionModal question toolInstant
Verified
Lifetime License · No Subscriptions

Ready to Arm Your Agent?

Get lifetime access to 50+ directory playbooks, automated relevance matching, and the verified 12-state submission machine.