Revora

Claude Code integration

claude code illustration for Revora

Endpoints

MCP
https://revora-ai.com/api/agent/v1/mcp
REST
https://revora-ai.com/api/agent/v1/*

Setup

  1. 1

    Create and store a token

    Generate a Revora API key and keep it in the environment Claude Code already uses for secrets.

  2. 2

    Call the hosted MCP server

    POST https://revora-ai.com/api/agent/v1/mcp with the Bearer token, or use the REST routes.

  3. 3

    Read before write

    Have Claude Code fetch context and list agents from the repo session before it creates a finder.

  4. 4

    Review in Overview

    After any write, open Revora and confirm the agent, limits, and targeting match the session.

Why Claude Code users connect Revora

Claude Code is built for agentic work in a repo and terminal. Connected to Revora, it can operate sales workspace tools while you keep reviewing diffs, logs, and approvals in a developer-native loop.

Setup

Create a Revora API key. Configure Claude Code's MCP or HTTP tooling with the MCP endpoint or REST base URL and Bearer auth. Point the agent at agents.md so it learns guardrails before it creates agents.

  • Prefer environment variables for secrets
  • Start with get_context and get_stats
  • Require explicit approval for create, delete, and reply
  • Do not commit API keys to the repository

Useful workflows

Refresh My Product from a README or landing page draft you trust, spin up a single test lead finder, dump lead JSON for offline review, or audit scheduled actions before a campaign weekend. Keep production changes small and reversible.

Claude Code versus Claude chat connector

Claude's chat product can use a custom connector with workspace approval. Claude Code typically uses a key like other coding agents. Pick the path that matches the client. Do not reuse chat OAuth assumptions in a headless runner.

Using Claude Code next to the product repo

Claude Code is a good operator when the same session is already editing prompts, docs, or scripts. Ask it to compare My Product with what the landing page claims. If those disagree, fix the profile before you let it launch outreach.

Do not let a long coding session casually 'also spin up a campaign.' Creating agents should be a named request, the same way merging a pull request is a named request.

When to stay in the browser

If you are not in a terminal already, Overview or a chat connector is faster. Claude Code earns its place when the work is happening in a repo. It is a slow way to click 'create campaign' if that is the only job.

Frequently asked questions

The hosted MCP endpoint is the same product surface. Auth differs: chat connectors use workspace approval UI, Claude Code should use a revocable API key unless you implement another supported flow.

Only through Revora tools under your campaign and safety settings. Your prompts should still require approval for replies and agent lifecycle changes.

https://revora-ai.com/agents.md is the operator guide. Adapt it for Claude Code instructions.

Connect your operator to a real sales workspace

Revora holds the LinkedIn connection and safety limits. Your AI configures agents and inspects results.

Get startedMCP overview

Prefer markdown? Claude Code integration.md