6sense
Ask an AI agent about 6sense account insights, buying signals and keyword trends.
Run Chatbase agents, sources, conversations and help-desk tickets from Claude.
Opens Agent Studio, where connecting is one click. The connector URL below works in any MCP client.
Paste it into any MCP client. Setup docs
The Chatbase MCP server connects Claude to a Chatbase workspace: the AI support agents, the sources they are trained on, their conversations and the help-desk tickets behind them. It can create and retrain agents, edit sources, and reply to customers on tickets. You choose which read, write and delete permissions to grant on Chatbase's consent screen.
Verified connector
Listed by Anthropic as a partner connector in its Connectors Directory.
Connection checked by Agentman on .
Anthropic states this reflects the level of review a connector received, not a security audit.
Tool names from Anthropic's directory listing. This server requires sign-in, so we could not read tool descriptions or parameter schemas.
chatbase_chat returns the whole response in one message.Because the approval was narrower than the task. Chatbase's documentation says a client only sees the tools its approved permissions cover. Ask the assistant to call chatbase_whoami to see the permissions in use, then revoke the app under Connected apps and reconnect with the permissions it needs. If a permission never appears on the consent screen, your own Chatbase role lacks it and an administrator must grant it.
Yes. Chatbase's documentation states the workspace needs the Basic plan or higher. Without a paid plan, and on trials that signed up with a personal email address, every tool except chatbase_whoami returns an error. The error text starts with the code SUBSCRIPTION_API_RESTRICTED_PLAN, so a failing first call usually means a plan problem rather than a broken connection.
Yes, if you approve helpdesk write access. Chatbase's tool reference marks chatbase_create_ticket_message as a sensitive write that sends a reply the customer receives, or an internal note only your team sees. Chatbase says a client should ask you before every sensitive write. Approve only the read permissions at sign-in if you want the assistant to read tickets without being able to answer them.
Because training runs in the background after the write. Chatbase's documentation says a new source reports untrained until it is live, then trained, or failed if it did not land. An edited source reports updated while it retrains. Ask the assistant to read the source again rather than assuming the create or update call finished the job when it returned.
Approve only the read permissions on the consent screen. Chatbase groups permissions by resource and verb, such as sources:read and sources:write, and lets you choose which to grant. Its documentation says delete and export permissions are shown but not selected by default. Chatbase also caps every app at your own role, so an app cannot delete what you cannot delete.
documentation URL https://docs.chatbase.co/developer-guides/mcp redirects here). Source for the tool reference and labels, permissions, sign-in, token lifetimes, limits, errors and troubleshooting. · retrieved 2026-11-18initialize POST to https://mcp.chatbase.co/api/mcp, RFC 9728 metadata at https://mcp.chatbase.co/.well-known/oauth-protected-resource/api/mcp, and authorization-server metadata at https://www.chatbase.co/.well-known/oauth-authorization-server (2026-11-18). · retrieved 2026-11-18docs/marketing/data/anthropic-directory-2026-11-17.json. · retrieved 2026-11-17Connect Chatbase once and your agents call these tools on their own: on a schedule, in a workflow, with nobody at the keyboard.