Bring RepScale to your AI agent.
Connect RepScale to Claude Desktop or any MCP-compatible agent. Your agent gets the same five sales-research tools you use in the web app, and the same daily quota. No extra setup. No separate billing. Revoke any agent any time from Settings.
Listed on the Official MCP Registry, Smithery, Glama, and mcp.so.
Why this matters
Your agent should have the same research tools you do.
Most agents handle prospect prep by pasting a LinkedIn URL into their context window and asking the model what it knows. That produces a generic summary that any rep could write blind.
RepScale gives your agent real research tools instead: a structured brief generator with live web data, a recommended sales angle, source citations, and confidence signals. The same tools a real rep uses in the RepScale web app, callable programmatically from inside any MCP-compatible agent.
Five tools
What your agent can do.
research_prospectGenerate a structured brief on a prospect.Your agent gets the same research surface a rep uses in the web app: company snapshot, recent business signals, leadership context, and contact-level insights. Briefs cache for 7 days per prospect, so an agent looping on the same name doesn't burn your quota.
get_recommended_angleSkip the synthesis step.Instead of a pile of facts, your agent gets the recommendation: the hook, the pain point, and the proof a seller needs to bring. Hands the agent something it can act on, not just describe.
explain_confidenceTells the agent when a brief is shaky.Every research brief has a confidence level. This tool explains why the level is what it is and what would raise it. Your agent can then decide whether to escalate, gather more, or proceed.
suggest_adjacent_researchHelps the agent know what to dig into next.Suggests follow-up research the agent might want to run based on what is already in the brief. Useful for multi-step agents that handle prospect prep end to end.
cite_sourceReturns the citation behind any claim.Hand it a claim from a brief and it returns the source: URL, retrieval date, and the quoted passage that supports the claim. Agents that double-check before sending get a real grounding signal.
Three steps
Add RepScale to Claude Desktop.
- 1
Drop the server URL into your client config
In Claude Desktop, open Settings → Developer →
Edit Config. Add this entry:{ "mcpServers": { "repscale": { "url": "https://app.repscale.ai/api/mcp" } } }Other MCP clients use a similar config; your client's docs will show the exact format. The server URL is the only required field.
- 2
Approve the agent in your browser
The first time the agent makes a call, it walks the OAuth flow. RepScale opens a browser tab showing which agent is connecting and what scope it wants. Click Approveand you're done. Standard OAuth 2.0 with PKCE and dynamic client registration. No API key to copy, paste, or store.
- 3
Start using the tools
Ask your agent to research a prospect, get a recommended angle, or pull a source citation. Calls count against your normal RepScale daily quota. Free tier covers 1 research call per day; Pro covers 200.
You stay in control.
- →OAuth 2.0 with PKCE for end-user agents. No API key sharing required.
- →API keys available for backend integrations. Hashed at rest, plaintext shown once at creation.
- →Every authorized agent appears in Settings → Connected Agents. One click to revoke. Revocation takes effect immediately.
- →Tokens last one hour. Refresh tokens rotate on every use. Reuse triggers a full revocation cascade.
- →Agents share the daily quota of the user who authorized them. No double billing. Free tier still works for agents.
- →Inputs flow to Anthropic and Perplexity (the same way the web app uses them). Neither retains data for training. See the privacy policy for the full data flow.
Ready to give your agent real sales tools?
Start with the free tier. No credit card. Connect Claude Desktop once and skip the busy work on every prospect after that.