MCP App Store
by KeenEthics
Salespeak Company Context icon

Salespeak Company Context MCP Server for Claude and ChatGPT

By Salespeak

Overview

Ask what your company has recorded about its product, pricing, integrations, security, customers and positioning, and get an answer with the records behind it. Use it before you write a prospect email, prepare for a call, answer a security question or draft a comparison, so the facts come from your company's own knowledge and not from a guess. When two of your company's sources disagree, a card shows both versions with the passage that states each one. You pick one for the answer you are writing, or flag the conflict to an admin. The connector cannot decide which version is right: an admin settles that in Salespeak, and every later answer follows the decision. When your company has no recorded answer, Claude says so and does not invent one. Three tools read your company's knowledge. One tool writes: it records a flag for your admins and changes no answer. Requires a Salespeak account. You sign in with it, and the connector reads only the knowledge of the company that account belongs to.

How to connect

  1. Claude: Open the Salespeak Company Context listing in the Claude connectors directory and choose Connect. Claude listing Salespeak Company Context exposes 4 tools here.
  2. ChatGPT: Open the Salespeak Company Context listing in the ChatGPT app directory and choose Connect. ChatGPT listing Salespeak Company Context exposes 4 tools here.

Tools

ask_company_context

ChatGPT
Answer a question about the user's own company from the knowledge it keeps in Salespeak. Use it when a request depends on a fact about the user's own company: its product and capabilities, pricing and plans, integrations, security and compliance, customers, competitors, ideal customer, or how it positions and describes itself. Typical requests are a prospect email, call prep, an objection, a comparison or a partner intro. It is not meant for rewriting or summarizing text the user provided, for tone changes, or for questions unrelated to the user's company. It reads the company's knowledge and changes nothing. Which version of a disputed fact is right is decided by an admin in the Salespeak app, not by this tool. Args: query: What you need to know about the company, as a question (required). One topic per call. Returns: JSON with: - status: "answered", "partial", or "not_in_kb" when the company has no recorded answer. - answer: the answer, grounded in the company's knowledge bank. - records: the records behind the answer, each with an authority (approved means a person at the company approved it); authority_summary, freshness and gaps describe them. - conflicts: where the company's own sources disagree on this question. The user sees these versions in a card and chooses one there. - other_open_conflicts: how many other open conflicts the company's knowledge holds. - knowledge_scan: whether the knowledge was ever checked for conflicts ("done", "running" or "never") and the page where an admin runs the check. - notes: plain statements about this result, for example that the user has not chosen between conflicting versions yet.

flag_knowledge_conflict

ChatGPT
Mark, in the user's own company workspace, that two of its records disagree and need a decision. Use this when the user says a company answer is wrong, or asks to report or escalate a disagreement between company sources. It stores one flag in the same Salespeak workspace this app reads from, the one the user signed in to: who raised it, and optionally which record they believe is right and why. It sends nothing to any other system, service or person, and changes no answer. The company's admins see the flag the next time they review conflicts in that workspace. Args: conflict_id: The conflict's id, from the conflict_id of a conflict in a company context answer or in the list of open conflicts (required). preferred_record_id: The conflict's record the user believes is right, if they said. note: A short reason in the user's words (up to 1000 characters). Returns: JSON confirming the flag, or an error message.

list_knowledge_conflicts

ChatGPT
List where your company's own knowledge contradicts itself. Use this tool to see the open conflicts between records in your knowledge bank, found by the offline knowledge bank scan: two sources that state different prices, limits, plans, dates or definitions for the same thing. Perfect for: - "What is inconsistent in our company context?" - "Where do our sources disagree?" - "What should product marketing settle first?" Args: limit: Maximum conflicts to return (default 10, capped at 25). Returns: JSON with conflicts, most severe first. Each has a stable conflict_id, the topic, how the records disagree, and both records with their sources.

quote_conflict_passages

ChatGPT
Quote, for each record of a conflict, the passage that states the disputed fact. It exists for the Company Context card, which calls it once a conflict is on screen so each option can show the sentence that disagrees and not the start of the record. It reads only records of the user's own company and changes nothing. Args: conflicts: Up to 5 items of {"record_ids": [...], "note": "..."}, copied from the conflicts of a company context answer (required). Returns: JSON {"claims": [{record_id: passage}, ...]}, one object per conflict, in order. A record with no such passage has no entry.

ask_company_context

Claude

flag_knowledge_conflict

Claude

list_knowledge_conflicts

Claude

quote_conflict_passages

Claude

Similar Apps

Listed on MCP App Store

Listed on MCP App Store

Add this badge to your site or README. It links to this page.

HTML

<a href="https://mcp-app-store.keenethics-labs.com/app/salespeak-company-context"><img src="https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg" alt="Listed on MCP App Store" width="220" height="44"></a>

Markdown

[![Listed on MCP App Store](https://mcp-app-store.keenethics-labs.com/badges/listed-on-mcp-app-store.svg)](https://mcp-app-store.keenethics-labs.com/app/salespeak-company-context)