Claude integration

Endpoints
- MCP
- https://revora-ai.com/api/agent/v1/mcp
- REST
- https://revora-ai.com/api/agent/v1/*
- Guide
- /agents.md
Setup
- 1
Finish Revora setup
Create an account, connect LinkedIn, and fill My Product so tools have something true to say.
- 2
Add the connector in Claude
Settings, Connectors, custom connector. Use https://revora-ai.com/api/agent/v1/mcp.
- 3
Approve the workspace
Sign in on Revora and allow Connect workspace. Enable tools in the conversation if Claude asks.
- 4
Read context first
Ask Claude to run get_context and get_stats, then list agents, before it creates anything.
What Claude can do with Revora
After Claude is connected to the Revora MCP server, it can use Revora tools to read workspace readiness, update My Product, create classic lead finders or Steal Customers agents, list leads with engagement context, check activity and the planned send schedule, and work with existing conversations under draft approval rules.
Claude does not receive your LinkedIn password. It operates through Revora's workspace tools, with the same daily safety limits and campaign rules as Overview.
Setup steps
Create a Revora account, connect LinkedIn, and complete My Product. In Claude, open Settings, go to Connectors, and add a custom connector with the Revora MCP URL. Sign in on Revora when prompted and approve Connect workspace. Enable Revora tools in the conversation if Claude asks.
- MCP URL: https://revora-ai.com/api/agent/v1/mcp
- No API key required for the chat connector path
- Approve workspace access on Revora before tools work
- Optional: point the agent at https://revora-ai.com/agents.md as the first instruction
Good first requests
Ask Claude to run get_context and get_stats, confirm My Product is complete, list existing agents before creating new ones, and only create a lead finder or Steal Customers agent after you explicitly approve the targeting. Never ask it to broaden ICP silently or reply to threads without showing you the draft.
Limits and responsibilities
You remain responsible for LinkedIn rules, the accounts you connect, and the offers you send. Claude is an operator interface. Revora enforces product limits; it does not replace your judgment about who to contact or what to claim.
A first Claude session that stays safe
Start with a question Claude can answer from tools: is My Product complete, how many agents exist, what is scheduled to send. Then give it one job, such as creating a lead finder for a written ICP, and make it show the targeting before it saves.
Do not ask Claude to invent a broader market or to reply to threads without showing the draft. The connector is powerful because it can change a real workspace. Treat it like a teammate with admin access, not like a chatbot that only talks.
When Claude is the wrong operator
If you need scripts, CI, or a coding agent in the repo, use an API key with Cursor, Claude Code, or REST. Claude's connector path is for conversational operation. It is a poor fit if you want headless jobs or if you are not willing to approve what it creates.
Frequently asked questions
Not for the custom MCP connector flow. Claude uses the connector URL and workspace approval. API keys are for coding agents and scripts.
Yes, after My Product is complete and LinkedIn is connected. Provide competitor LinkedIn company URLs and optional founder or employee profile URLs.
See the MCP integration page for the tool groups, and agents.md for the machine guide.
Connect your operator to a real sales workspace
Revora holds the LinkedIn connection and safety limits. Your AI configures agents and inspects results.
Prefer markdown? Claude integration.md