AgentStack
Browse Sign in
Browse Why AgentStack Sell Docs
Sign in
MCP verified MIT Self-run

Vantage

mcp-cavendo-ai-vantage · by Cavendo-AI

Capture and query what you’re seeing: a shared memory layer for AI-native work

No reviews yet
0 installs
30 views
0.0% view→install

Install

$ agentstack add mcp-cavendo-ai-vantage

✓ scanned · ✓ verified, works with Claude Code, Cursor, and more.

Security review

✓ Passed

No issues found. Passed automated security review. · v0.1.1 How review works →

  • Prompt-injection patterns
  • Secret / credential exfiltration
  • Dangerous shell & filesystem operations
  • Untrusted network calls
  • Known-malicious package signatures

What it can access

  • Network access Used
  • Filesystem access No
  • Shell / process execution No
  • Environment & secrets No
  • Dynamic code execution No

From automated source analysis of v0.1.1. “Used” means the capability is present in the source — more access means more to trust, not that it’s unsafe.

View the full security report →

Verified badge

Passed review? Show it. Paste this badge into your README, it links to the public security report.

AgentStack Verified badge Links to your public security report.
[![AgentStack Verified](https://agentstack.voostack.com/badges/verified.svg)](https://agentstack.voostack.com/security/report/mcp-cavendo-ai-vantage)

Reliability & compatibility

Security review passed
0 installs to date
no reviews yet
6mo ago

Declared compatibility

Claude CodeClaude DesktopCursorWindsurf

Compatibility is declared by the source manifest. End-to-end runtime verification is coming, see below.

Preview Execution monitoring

We're building live execution health for every listing: tool-call success rate, median latency, uptime, and last-checked timestamps, measured, not self-reported. It isn't live yet, so we don't show numbers we can't stand behind.

How agent discovery & health will work →
Are you the author of Vantage? Claim this listing to set pricing, connect Stripe payouts, and keep 70% of every sale.
Sign up to claim

About

Vantage

Track what you're seeing. Use it later.

Vantage is a personal signal store for AI-native work — a place to capture and query the things you notice across the web.

A post on X. A thread on LinkedIn. An insight from an article. A competitor move. Instead of losing it across tabs, screenshots, and notes, you save it — and use it later.


One memory layer across all your AI tools

Your AI tools don't share memory.

  • Claude doesn't know what you saw on X
  • ChatGPT doesn't know what you read yesterday
  • Your code tools don't know your research

Vantage fixes that.

Save signals once, and access them anywhere you have an MCP-compatible client — across Claude, Cursor, and other supported tools.


How it works

  1. Connect Vantage to your AI (via MCP)
  2. Save signals directly from your workflow
  3. Ask questions on top of what you've saved

Example:

> "Save this as a signal"

> "What have I saved about pricing?"

> "Show me signals from Lemkin"


What Vantage is (and isn't)

  • Not a scraper — you choose what matters
  • Not a feed — no noise or algorithmic clutter
  • Not a knowledge base — no system to maintain

Vantage is a simple way to keep track of what you're seeing — so your AI can actually use it later.


Hosted + open source


Quick Start

Option 1: Use the hosted version (fastest)

Sign up at vantage.cavendo.ai. Connect to Claude or another supported client. Start capturing.

Takes about 2 minutes. No install required.

Option 2: Run locally

npm install
npm run dev

# If 3020 is already in use, choose another port
PORT=YOUR_PORT npm run dev

# Generate your first API key
curl -X POST http://localhost:3020/api/auth/keys \
  -H 'Content-Type: application/json' \
  -d '{"name": "my-key"}'

The first API key can be created from localhost without auth. If you're bootstrapping a fresh remote deployment, set VANTAGE_BOOTSTRAP_TOKEN on the server and send the same value in the X-Vantage-Bootstrap-Token header on that first request. The API key is returned as data.key in the JSON response.

Capture your first signal:

curl -X POST http://localhost:3020/api/signals \
  -H 'Content-Type: application/json' \
  -H 'Authorization: Bearer vtg_YOUR_KEY_HERE' \
  -d '{
    "signalType": "post",
    "platform": "x",
    "content": "Every SaaS company will need an agent layer within 18 months.",
    "topics": ["AI Agents", "SaaS"],
    "importance": "high"
  }'

MCP Server

Vantage includes an MCP server with 13 tools for use with Claude, Claude Code, Cursor, or any MCP-compatible client.

If you want to use the bundled stdio MCP server directly, install its dependencies too:

cd mcp-server
npm install
cd ..

Claude Code

claude mcp add vantage --transport http "http://localhost:3020/mcp" --header "Authorization: Bearer vtg_YOUR_KEY"

For manual testing, note that /mcp is a streamable HTTP endpoint. MCP clients handle the response format automatically, but ad hoc curl calls need to be prepared for stream-oriented responses rather than a simple JSON-only request/response flow.

Claude Desktop

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

{
  "mcpServers": {
    "vantage": {
      "command": "node",
      "args": ["/path/to/vantage/mcp-server/src/index.js"],
      "env": {
        "VANTAGE_API_URL": "http://localhost:3020",
        "VANTAGE_API_KEY": "vtg_YOUR_KEY_HERE"
      }
    }
  }
}

MCP Tools

| Tool | What it does | |------|-------------| | capture_signal | Save a post, article, quote, or discussion. Auto-creates sources and topics. | | search_signals | Full-text search with filters — platform, topic, date range, importance. | | list_recent_signals | Latest signals feed. | | get_signal | Full detail on a specific signal. | | manage_source | Create or update a person/account you're tracking. | | list_sources | See everyone you're tracking with signal counts. | | manage_topic | Create or update a topic/theme. | | list_topics | All topics with signal counts. | | manage_collection | Group signals into named collections. | | set_business_context | Upload strategy docs for comparison. | | analyze_signals | Assemble signals + context for analysis. | | get_dashboard | Summary — counts, top topics, top sources, highlights. | | save_analysis | Persist an analysis for later reference. |


Integrations

  • OpenClaw community skill: [vantage-shared-intel](./integrations/openclaw/vantage-shared-intel/SKILL.md) — capture signals, search prior intel, and generate shared digests against the same Vantage dataset from OpenClaw.

REST API

Most API requests require Authorization: Bearer vtg_YOUR_KEY.

Exceptions:

  • GET /health
  • POST /api/auth/keys for first-key bootstrap from localhost, or from a remote client that supplies a valid X-Vantage-Bootstrap-Token

| Method | Endpoint | Description | |--------|----------|-------------| | POST | /api/signals | Create a signal | | GET | /api/signals | List signals with filters | | GET | /api/signals/search?q=query | Full-text search | | GET | /api/signals/feed | Chronological feed | | GET | /api/signals/:id | Signal detail | | POST | /api/sources | Create a source | | GET | /api/sources | List sources | | POST | /api/topics | Create a topic | | GET | /api/topics | List topics | | POST | /api/collections | Create a collection | | POST | /api/contexts | Upload business context | | GET | /api/dashboard/summary | Dashboard summary | | POST | /api/auth/keys | Generate API key |


Tech Stack

| Component | Technology | |-----------|-----------| | Runtime | Node.js 20+ / Express / ESM | | Database | SQLite via better-sqlite3 | | Validation | Zod | | Auth | API key (vtg_ prefix), SHA-256 hashed | | MCP | @modelcontextprotocol/sdk (stdio + HTTP) |


Environment Variables

| Variable | Default | Description | |----------|---------|-------------| | PORT | 3020 | Server port | | NODE_ENV | development | Environment | | DATABASE_PATH | ./data/vantage.db | SQLite file path | | CORS_ORIGIN | * | CORS allowed origin | | VANTAGE_BOOTSTRAP_TOKEN | unset | Allows first API key creation from a non-localhost client when sent as X-Vantage-Bootstrap-Token |


License

MIT — use it however you want.


Built by Jonathan Arehart at Cavendo AI.

Source & license

This open-source MCP server is cataloged on AgentStack and links to its original source — we do not rehost the code.

Install and usage instructions live in the source repository linked above.

Reviews

No reviews yet, be the first.

Versions

  • v0.1.1 Imported from the upstream source.