Centillion
Remote MCP server for building and testing Centillion AI agents from chat clients
- Pricing
- Contact sales
- Platforms
- Web, API
- Runs
- Remotely
- Version
- 1.0.0
- Listing updated
- 2 days ago
About Centillion
Centillion is an MCP server for ai models and llm tooling. It runs as a hosted remote server, so there is nothing to install locally.
Centillion's server lets Claude, ChatGPT, Cursor or another MCP client work inside a Centillion workspace: look up agents, edit them and test them from a conversation. Anything that would go live needs a person to approve it first, through request tools such as request_publish, request_traffic_change and request_batch_launch.
Access follows workspace roles. Owners and Members can use every tool, Viewers can use only the reading tools, and Support Associates are refused. Changes on Main or on a branch carrying traffic land in the user's own draft, while a quiet branch gets the change as its next version straight away.
The read side covers agents, branches, workflows, voice settings, tools, batch calls, phone numbers, knowledge sources, conversations and test results. Secrets such as header values and OAuth client secrets are never returned. A Claude plugin is also offered.
Install
Run in a terminal
claude mcp add --transport http centillion https://app.cntillion.com/mcpAdd to Settings → Connectors → Add custom connector
https://app.cntillion.com/mcpAdd to ~/.cursor/mcp.json
{
"mcpServers": {
"centillion": {
"url": "https://app.cntillion.com/mcp"
}
}
}Add to .vscode/mcp.json
{
"servers": {
"centillion": {
"type": "http",
"url": "https://app.cntillion.com/mcp"
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"centillion": {
"serverUrl": "https://app.cntillion.com/mcp"
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"centillion": {
"type": "http",
"url": "https://app.cntillion.com/mcp"
}
}
}Tools this server provides (33)
Taken from the project documentation; the running server may differ.
- list_agents
- Lists the workspace's agents with id, name, role, model and training status.
- get_agent
- Returns one agent's setup as last published on Main.
- list_branches
- Lists an agent's branches.
- get_workflow
- Returns the workflow on a branch's current version.
- get_voice_settings
- Returns the voice settings a branch's calls use.
- list_tools
- Lists workspace tools without secrets.
- list_batch_calls
- Lists outbound call batches.
- list_phone_numbers
- Lists workspace phone numbers.
- get_message_settings
- Returns message templates, opt-out lines and sender status.
- list_knowledge_sources
- Lists an agent's knowledge sources.
- list_conversations
- Lists an agent's conversations.
- get_conversation
- Returns one conversation's messages.
- get_test_results
- Returns an agent's latest test runs.
- search_docs
- Searches the help pages.
- create_agent
- Creates an agent with a Main branch.
- update_agent_draft
- Changes a draft of an agent's settings.
- create_branch
- Starts a branch.
- save_workflow
- Replaces a branch's workflow.
- update_voice_settings
- Changes a branch's voice settings.
- create_tool
- Creates a workspace tool.
- update_tool
- Changes a tool's settings.
- discover_mcp_tools
- Asks an MCP server tool which tools it offers.
- save_test
- Creates or replaces an agent test in a draft.
- run_tests
- Runs tests on a draft.
- save_message_template
- Creates or replaces a message template.
- create_connector_action
- Creates a connector action tool.
- add_knowledge_text
- Adds a text snippet to knowledge.
- add_knowledge_website
- Imports a public web page into knowledge.
- create_email_sender
- Gives an agent an email address.
- request_publish
- Asks for approval to publish a draft.
- request_traffic_change
- Asks for approval to split conversations between branches.
- request_batch_launch
- Asks for approval to launch a call batch.
- request_archive
- Asks for approval to archive a branch or tool.
What it does
- Agent lookup
- Lists agents and returns one agent's instructions, model, guardrails, workflow and tests.
- Draft editing
- Changes instructions, greetings, models, workflows and voice settings in a personal draft.
- Test runs
- Saves tests and runs them on a draft, with results readable afterwards.
- Approval gates
- Publishing, traffic splits, batch launches and archiving wait for a person.
- Knowledge management
- Adds text snippets or a public web page to an agent's knowledge.
- Role-based access
- Viewers read only; Owners and Members use every tool.
Good for
- Draft and test a customer support agent from Claude or Cursor
- Review an agent's workflow and guardrails without opening the dashboard
- Check recent test results before asking for a publish
- Add knowledge text or a web page to an agent
- Inspect phone numbers and batch calls attached to a workspace
Pricing
Contact sales
A paid Centillion subscription is required; contact the team for plans
Prices change. Check cntillion.com before you buy.
Centillion FAQ
Where do I connect?
The server address is https://app.cntillion.com/mcp, added to the MCP client's server list.
Can an assistant publish changes by itself?
No. Publishing waits for a person to approve the request.
Who can use the writing tools?
Owners and Members; Viewers can only read.
Are secrets exposed?
Header values and OAuth client secrets are never returned.
Is there a Claude plugin?
Yes, installed from the Centillion-Labs marketplace entry.
How much does Centillion cost?
Centillion does not publish prices. Contact the company for a quote.
How do I add Centillion to Claude Code?
Run this command in your terminal: claude mcp add --transport http centillion https://app.cntillion.com/mcp
What are the best alternatives to Centillion?
The closest alternatives to Centillion are REA, Codebase Memory, GitHub, Serena MCP: the IDE for your agent and XERJ.
Alternatives to Centillion

REA
MCP serverCLI and MCP server that lets coding agents reverse engineer apps and binaries
Runs locally- Open source
- 464K GitHub stars
- Developer tools

Codebase Memory
MCP serverLocal MCP server that builds a code knowledge graph for AI coding agents
Runs locally- Open source
- 48.7K GitHub stars
- Knowledge and memory

GitHub
MCP serverOfficial MCP server connecting AI assistants to GitHub repos, issues and workflows
Not checked yet- Remote
- Open source
- 34.5K GitHub stars
- Developer tools

Serena MCP: the IDE for your agent
MCP serverMCP toolkit giving coding agents symbol-level code retrieval and editing
Runs locally- Open source
- 31.5K GitHub stars
- Developer tools

XERJ
MCP serverLocal search engine for AI agents with cited answers over your code and docs
Runs locally- Open source
- 19.2K GitHub stars
- Knowledge and memory

Ouroboros
MCP serverSpec-first layer that pins requirements and verifies results of AI coding agent work
Runs locally- Open source
- 6.4K GitHub stars
- Developer tools