Gorilla MCP
This MCP helps you find your first 100 users on social media channels using Gorilla, a tool designed for founders by founders.
README
Gorilla MCP Server
Model Context Protocol server for Gorilla. Find your first 100 SaaS users by searching Reddit, YouTube, X, and TikTok for real demand signals.
Learn more: usegorilla.app · Real run examples · How Gorilla compares · Blog
<a href="https://glama.ai/mcp/servers/opusforge/gorilla-mcp"> <img width="380" src="https://glama.ai/mcp/servers/opusforge/gorilla-mcp/badges/card.svg" alt="gorilla-mcp MCP server"> </a>
Setup
Run directly from GitHub with npx:
GORILLA_API_KEY=grla_... npx -y github:opusforge/gorilla-mcp
Or clone and run locally:
git clone https://github.com/opusforge/gorilla-mcp
cd gorilla-mcp
npm install && npm run build
GORILLA_API_KEY=grla_... node dist/index.js
Get your API key
- Sign up at usegorilla.app and pick a plan ($0.99 single run, $3.99/wk Pro, or $149.99 lifetime).
- Sign in at gorilla.opusforge.com.br.
- Menu, API Keys, Create.
- Copy the key (shown once).
Configure in Claude Code
Add to your Claude Code settings:
{
"mcpServers": {
"gorilla": {
"command": "npx",
"args": ["-y", "github:opusforge/gorilla-mcp"],
"env": {
"GORILLA_API_KEY": "grla_your_key_here"
}
}
}
}
Configure in Cursor
Add to .cursor/mcp.json:
{
"mcpServers": {
"gorilla": {
"command": "npx",
"args": ["-y", "github:opusforge/gorilla-mcp"],
"env": {
"GORILLA_API_KEY": "grla_your_key_here"
}
}
}
}
Tools
Tools are namespaced by domain (leads.*, idea.*, runs.*, outreach.*, account.*).
leads.find
Run the full pipeline. Searches Reddit, X, YouTube, and TikTok and returns scored leads. Takes 60-120 seconds. Costs 1 run credit.
Parameters: idea (required) — product description
Returns: Scored leads with source, channel, title, URL, lead_score (0-1), and outreach hints.
idea.refine
Conversational refinement. Returns one clarifying question at a time to sharpen the idea before searching. Free.
Parameters: idea (required), plus current_refined_idea, history, language, turn, max_turns (all optional)
Returns: Status, refined_idea, readiness_score, and the next question (or null when ready).
idea.expand
Generate keyword scaffolding (core keywords, adjacent niches, pain points, competitor names, exclusion terms) without running searches. Costs 1 run credit.
Parameters: idea (required)
Returns: Structured themes for targeted searches.
leads.search
Search a single platform with custom queries. Bypasses theme expansion and AI scoring. Costs 1 run credit.
Parameters:
source(required):reddit,x,youtube, ortiktokqueries(required): Array of search queriesrun_id(optional): Attach results to an existing run
Returns: Raw leads from the specified platform.
runs.get
Fetch results for a previously-started run. Free.
Parameters: run_id (required)
runs.list
List your last 50 runs, newest first. Free.
account.billing
Check your plan, remaining weekly runs, and referral credits. Free.
Returns: Plan name, weekly usage, referral credits, total available runs.
outreach.draft
Generate a platform-tuned outreach message for a specific lead. Costs 1 run credit per draft.
Parameters: idea, source, outreach_action, post_title, post_body (required), plus optional post_handle, language, reply_to_author, reply_to_text.
Returns: A ready-to-paste draft.
outreach.plan
Build a Week-1 outreach plan from a completed run's HIGH-intent leads, with per-channel send cadence. Free.
Parameters: run_id (required)
Example workflow
1. idea.refine("a language learning app for travelers")
→ "Who's the target user? Daily commuters or tourists?"
2. leads.find(refined_idea)
→ 47 leads across Reddit, YouTube, X, and TikTok
→ 12 high-intent (people actively searching for this)
3. outreach.plan(run_id)
→ Week-1 plan: 3/day on Reddit, 4/day on X, 2/day on YT/TT
4. outreach.draft(...) → ready-to-send reply for each high-intent lead
Install via Smithery
Available at smithery.ai/server/opusforge/gorilla-mcp. Smithery distributes a pre-built MCPB bundle that any MCPB-compatible client can install in one click.
Environment variables
| Variable | Required | Description |
|---|---|---|
GORILLA_API_KEY |
Yes | Your API key (starts with grla_) |
GORILLA_DEFAULT_LANGUAGE |
No | Fallback language for idea.refine and outreach.draft (en, pt, all). Default: en. |
GORILLA_CONFIG_URL |
No | Override the runtime config endpoint. Useful for staging or self-hosted deployments. |
Backend URL and gateway key are fetched automatically from https://gorilla.opusforge.com.br/mcp-config.json on startup. No other configuration required.
Pricing
- Single run: $0.99. Pay per use, no subscription.
- Weekly Pro: $3.99/week, 5 runs.
- Lifetime: $149.99 once, unlimited runs.
leads.find, leads.search, idea.expand, and outreach.draft each cost 1 run credit. idea.refine, runs.get, runs.list, account.billing, and outreach.plan are free.
See usegorilla.app for current plans and the full product.
推荐服务器
Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。
VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。