M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
Y

YALC

Community
Not a standard server
Open-source, AI-native GTM operating system. CLI-first.

This is not a typical connect-and-use MCP server; read the source repository before using it.

How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #258 of 438
Stars
★ 318 Popular
Transport
stdio (local process)
Runtime
Node.js 20+
Credentials
Optional API key
License
MIT
Last commit
Tools
20
50FMRS · D

YALC provides a powerful CLI framework for GTM automation with a rich set of skills and provider integrations. It is ideal for iterative sales teams comfortable with the command line, but requires careful setup and awareness of external API dependencies.

Strongest · Documentation 13/20 Weakest · Reliability 5/20

Reliability
5/20
Security and permissions
10/20
Maintenance
12/20
Documentation
13/20
Setup experience
10/20
Why each score
Reliability 5/20
Evidence is limited to the untrusted README: the project is explicitly a CLI ('no public Node API to import') with no MCP server manifest, tool list, or init handshake in evidence ('SERVER MANIFEST' not cached); MCP integration is described as 'current status, and roadmap' rather than a shipped capability. A large command surface is declared, but several commands are designed to redirect and exit cleanly without doing work, and no tests or CI workflow contents are verifiable. Static calibration caps reliability at 12 without real CI plus committed tests; the fundamental mismatch of reviewing this as an MCP server further lowers the score to 5.
Security and permissions 10/20
No red-line issues found: no malware, no real secrets in install examples (only sk-ant-... placeholders), no covert exfiltration or fabricated identity. Main risks are visible: many third-party API keys, outbound email/LinkedIn DM sends, a Vercel deploy adapter, and local storage under ~/.gtm-os with ENCRYPTION_KEY. However, least-privilege design, confirmation gates for dangerous operations, and data boundaries cannot be verified from supplied material. The README states positive practices: outbound validation, no silent mocks, and never logging API keys. Score 10.
Maintenance 12/20
MIT license, repo not archived, only 7 open issues, README references v0.13.0 and a CHANGELOG, and a CI badge is shown. But commit history, release cadence, issue response behavior, and dependency updates are not verifiable from the supplied evidence (README is untrusted), leaving governance/versioning gaps. Score 12.
Documentation 13/20
As a CLI, the README is layered and extensive: quick start, full command list, providers, skills, config with env precedence, plus links to first-run, providers, commands, skills, MCP, troubleshooting, and background-agent docs, with limitations disclosed (optional Playwright, optional key modes). But as an MCP server it lacks client configuration examples and tool parameters, docs/mcp.md is only described as 'current status, and roadmap', and linked doc contents cannot be independently verified. Score 13.
Setup experience 10/20
The CLI install path is genuinely few-step: npm global install then yalc-gtm start, with interactive, --review-in-chat, and --non-interactive (CI) modes, and a clear Node 20+ requirement. However, no mcpServers configuration example for any mainstream client is provided, and the static calibration caps setup at 15 without verifiable CI plus committed tests; scored 10.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessWrites / deletes local filesRuns commands or codeUses the networkConnects to a database

Best for

  • SMB B2B teams looking for an open-source, customizable GTM automation tool.
  • Developers who use Claude Code and want to integrate GTM workflows.
  • Operations teams that prefer CLI or MCP-controlled end-to-end campaigns.

Not for

  • Enterprise users needing a fully managed, GUI-driven GTM platform.
  • Teams requiring a pre-built, maintained MCP server rather than a CLI tool.
  • Users seeking official support from a CRM/ESP or similar vendor.

Required permissions

  • Access external APIs (Unipile, Crustdata, Firecrawl, Notion, FullEnrich, Instantly, Anthropic, etc.)
  • Read/write local filesystem (`~/.gtm-os/`, `./data/`)
  • Start a local dashboard server (port 3847) and open browser
  • Run background agents on your system (launchd)

Risks and side effects

  • May send unsolicited messages if misconfigured, potentially violating platform policies.
  • API keys stored locally in `~/.gtm-os/.env`; exposure could lead to unauthorized usage.
  • Automated messages may not always align with brand voice; human review is recommended.
  • Relies on multiple external services that may change APIs or incur costs.

Setup

Before you start

Runtime：Node.js 20+

ANTHROPIC_API_KEY optionalsecret Anthropic API key for LLM planning/qualification/personalization when running standalone (optional inside Claude Code); get from console.anthropic.com.
UNIPILE_API_KEY optionalsecret Unipile API key for LinkedIn search, connections, DMs and scraping; from Unipile dashboard.
NOTION_API_KEY optionalsecret Notion integration token for database sync; from Notion integrations settings.
ENCRYPTION_KEY optionalsecret Encryption key for stored credentials; generate with 'openssl rand -hex 32'.
FIRECRAWL_API_KEY optionalsecret Firecrawl API key for web scraping/search (optional inside Claude Code, which provides WebFetch); from firecrawl.dev.
CRUSTDATA_API_KEY optionalsecret Crustdata API key for company/people search and enrichment; from Crustdata.
FULLENRICH_API_KEY optionalsecret FullEnrich API key for email/phone enrichment; from FullEnrich.
INSTANTLY_API_KEY optionalsecret Instantly API key for cold email sending and sequence management; from Instantly.
VOYAGE_API_KEY optionalsecret Voyage API key for embeddings in the intelligence memory store; from Voyage AI.
LEMLIST_API_KEY optionalsecret Lemlist API key for the bundled lemlist skills/MCP integration; from Lemlist (see CLAUDE.md).
PEOPLEDATALABS_API_KEY optionalsecret API key for the bundled peopledatalabs people-enrich declarative adapter; from Peopledatalabs.
HUBSPOT_API_KEY optionalsecret API key for the bundled HubSpot crm-contact-upsert declarative adapter; from HubSpot.
BREVO_API_KEY optionalsecret API key for the bundled Brevo email-campaign-create declarative adapter; from Brevo.
VERCEL_TOKEN optionalsecret Vercel token for the bundled landing-page-deploy declarative adapter; from Vercel account settings.
Other optional settings (3)
UNIPILE_DSN optional Unipile DSN, formatted https://api{N}.unipile.com:{PORT}; from Unipile dashboard.
WEB_FETCH_PROVIDER optional Web-fetch provider override in .env.local: auto (default), firecrawl, or claude-code; not a secret.
VERCEL_TEAM_ID optional Optional Vercel team ID for the landing-page-deploy adapter; from Vercel.

Install globally: npm i -g yalc-gtm-os, then run yalc-gtm start. Requires Node.js 20+. Alternatively, open the repo in your IDE (Cursor/VS Code with Claude Code extension) and use the chat interface.

Check that it works

Run 'npm i -g yalc-gtm-os && yalc-gtm start' (or 'pnpm cli start' in-repo); success is confirmed when ~/.gtm-os/ is scaffolded and the dashboard opens at http://localhost:3847/setup/review, or by running 'yalc-gtm --help' and seeing the command list.

Troubleshooting

  1. Run `yalc-gtm setup` to check API keys and provider connectivity.
  2. Inspect `~/.gtm-os/config.yaml` for configuration settings.
  3. Ensure all required environment variables are set (e.g., `ANTHROPIC_API_KEY`).
  4. Check logs in `~/.gtm-os/logs/agents/` to debug background agents.

Things to try

Once connected, you can ask your AI assistant things like:

  • Run yalc-gtm leads:qualify --source csv --input ./leads.csv --dry-run on my lead list and show the qualification results.
  • Create a campaign called Q2 Outbound with the hypothesis that VP Eng responds to pain-point messaging.
  • Generate a weekly campaign intelligence report with campaign:report.
  • Sync my SQLite campaign data to Notion with notion:sync.

Tools 20

start writes
Guided onboarding: keys, context, framework, goals in one flow.
setup read-only
Check API keys and provider connectivity.
onboard writes
Build GTM framework from profile/website.
campaign:track writes
Poll Unipile, advance sequences, sync Notion.
campaign:create writes
Create campaign with A/B variant testing.
campaign:report read-only
Generate weekly intelligence report.
campaign:monthly-report read-only
Cross-campaign monthly report.
campaign:dashboard read-only
Open visualization dashboard.
Show 12 more tools
leads:qualify writes
Run 7-gate qualification pipeline.
leads:scrape-post read-only
Scrape LinkedIn post engagers.
leads:import writes
Import leads from CSV/JSON/Notion.
linkedin:answer-comments writes
Reply to LinkedIn post comments.
email:send destructive
Send a sequence or single message (choose provider).
email:create-sequence writes
Generate email drip sequence.
notion:sync writes
Bidirectional SQLite ↔ Notion sync.
notion:bootstrap writes
Import existing Notion data to SQLite.
orchestrate writes
Natural language → phased skill execution.
agent:run writes
Run background agent immediately.
agent:install writes
Install agent as launchd service.
agent:list read-only
List agents with last run status.

Use cases

Run multi-channel campaigns (LinkedIn + email) with automated tracking.
Import leads from CSV/JSON, qualify and enrich them.
Describe your GTM goal in natural language and let YALC plan the workflow.
Analyze campaign performance via dashboards and the intelligence store.

Supported clients

Claude Desktop

Listed from the project's documentation, not tested by this site.

Overview

YALC is an open-source, AI-native operating system for running any GTM campaign. It is CLI-first and features built-in skills, provider integrations, and an intelligence store. It serves as an open-source alternative to Clay, enabling multi-channel campaigns (LinkedIn + email), A/B variant testing, statistical significance, real-time dashboards, and DB-backed rate limiting.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 62.9k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 5686d1f2f526 Data synced 2026-10-12 Read the FMRS scoring method