Documentation Index
Fetch the complete documentation index at: https://docs.ando.so/llms.txt
Use this file to discover all available pages before exploring further.
To have your coding agent install Ando MCP for you, paste these instructions into the agent:
Install Ando MCP for me.
Follow these steps:
1. Ask me for my Ando API key before running any install command. Tell me I can get one by opening Ando, going to Settings -> API keys, then creating or copying an API key.
2. Ask which coding agent I want to install Ando MCP in. Supported choices are Claude Code or Codex CLI.
3. After I provide both the API key and the coding agent, run the necessary command for that agent.
Use these install commands:
Claude Code:
claude mcp add --transport http ando https://mcp.ando.so/mcp --header "Authorization: Bearer <ANDO_API_KEY>"
Codex CLI:
export ANDO_API_KEY="<ANDO_API_KEY>"
codex mcp add ando --url https://mcp.ando.so/mcp --bearer-token-env-var ANDO_API_KEY
Replace <ANDO_API_KEY> with the key I provide.
What is Ando MCP?
Ando MCP is Ando’s hosted Model Context Protocol server. It lets MCP-capable coding agents search and fetch context from your Ando workspace at:
https://mcp.ando.so/mcp
The server uses streamable HTTP MCP and API-key authentication.
Ando MCP includes public tools for conversations, messages, members, calls/Jams, clipboard items, and workspace actions. It supports both read actions like search, list, and fetch and explicit write actions like creating conversations, managing conversation members, sending messages, replying in threads, sending DMs, reacting to messages, deleting messages, and inviting workspace members.
Supported Clients
Ando MCP currently supports:
- Claude Code
- Codex CLI
- Cursor
Other MCP clients can connect if they support remote streamable HTTP servers with custom request headers.
Setup
First, create an Ando API key:
- Open Ando.
- Go to Settings -> API keys.
- Create or copy an API key.
- Use it anywhere the examples below show
ando_sk_... or ANDO_API_KEY.
Ando expects the key in a bearer token header:
Authorization: Bearer ando_sk_...
Claude Code
Add Ando MCP from your project directory:
claude mcp add --transport http ando https://mcp.ando.so/mcp \
--header "Authorization: Bearer ando_sk_..."
/mcp inside Claude Code and confirm that ando is listed.
Codex CLI
Store your Ando API key in an environment variable:
export ANDO_API_KEY="ando_sk_..."
codex mcp add ando --url https://mcp.ando.so/mcp --bearer-token-env-var ANDO_API_KEY
ANDO_API_KEY when it connects to the MCP server.
Cursor
Set the ANDO_API_KEY environment variable before launching Cursor:
export ANDO_API_KEY="ando_sk_..."
open -a Cursor
.cursor/mcp.json in your project or to ~/.cursor/mcp.json globally:
{
"mcpServers": {
"ando": {
"url": "https://mcp.ando.so/mcp",
"headers": {
"Authorization": "Bearer ando_sk_..."
}
}
}
}
ando is enabled.
Common Usage Patterns
Find a person, then browse their messages: Call search_members with a name or email, or list_members when you need active workspace members by exact display name. Take the returned member ID. Call get_conversation_messages with author to browse recent messages in a known conversation, or call search_messages with author if you need search instead of browsing.
Inspect a channel or DM roster: Call search_conversations or list_conversations to identify the conversation, then call list_conversation_members with the conversation ID.
Search for a topic, then inspect the full thread: Call search_messages with a keyword or semantic query. Inspect the returned replies_count. If the message is a thread root or has replies, call get_thread_replies with that message ID. If needed, call get_message for the exact source message.
Find meetings about a topic, then read the transcript: Call list_calls for recent calls/Jams or search_calls for transcript search. Pick a matching call_id. Call get_call for metadata. Call get_call_transcript to page through the transcript.
Find or update a task: Call search_tasks to discover visible tasks, then get_task for the task summary, recent work log entries, and active task URLs. Use record_task_update only when the user explicitly asks you to update a task.
Send a user-requested DM: Call search_members for each recipient unless IDs are already known. Then call send_direct_message with member_ids and markdown_content. Report the returned message_id and conversation_id.
React to a user-specified message: Confirm the exact target with get_message unless the user already provided a precise message ID, then call react_to_message with message_id and emoji.
Invite a user to the workspace: Call invite_to_workspace only when the user explicitly asks to invite an email address.
Start broad, then narrow: The usual workflow is to search or list conversations, members, messages, clipboards, calls, or tasks, capture the relevant IDs, then use the matching get_* or paginated tool for full details.
Behavior
Results are permission-aware and scoped to the authenticated member. search_messages supports "full-text" and "semantic" modes, defaults to "full-text", and rejects empty or wildcard-only queries. Use get_conversation_messages for recent history or author-filtered browsing instead of wildcard search.
get_conversation_messages returns messages in reverse chronological order. get_thread_replies returns replies in chronological order. get_call_transcript is paginated and should be called repeatedly with the returned cursor when needed. Use list_calls for recent calls/Jams without a search query, and search_calls when searching transcript content. Use list_conversations, list_members, and list_conversation_members for browse-style discovery without a fuzzy search query.
Write tools are side-effecting. Agents should only call them when the user explicitly asks to send, reply, DM, react, invite, update a task, or adjust proactivity. Before writing messages or reactions, confirm the target conversation, member, thread, or message from search or fetched data unless the user already provided a precise ID.
MCP sessions are stateful and idle-evicted after about 10 minutes, so clients should be prepared to re-initialize a session.
Recommended Prompting For Agents
Ando MCP works best when agents use search_* and list_* tools to discover the right IDs first, use get_conversation_messages instead of forcing search queries for recent history, use get_thread_replies when a message has replies, use list_calls for recent calls/Jams, use get_call_transcript only after identifying the right call, and avoid wildcard search inputs.
For write actions, prompt agents to verify the target before calling side-effecting tools. Examples: identify the conversation before send_message, identify the thread root before reply_to_message, identify recipients before send_direct_message, identify the message before react_to_message, confirm the email before invite_to_workspace, and fetch the task before record_task_update.
Troubleshooting
Unauthorized - API key required: Your client is not sending a valid bearer token. Check that the Authorization header is present, the value starts with Bearer , and the key has not been rotated or regenerated.
Cursor cannot see ANDO_API_KEY: Environment variables must be available to the Cursor app process. Launch Cursor from a terminal after exporting the variable, or paste the bearer token directly into .cursor/mcp.json.
Session errors: Ando MCP uses session-based streamable HTTP transport. If a client leaves a session idle for too long, it may need to re-initialize and reconnect.
Empty search results: Check that the authenticated member has access to the target workspace content, your query is specific enough to match indexed content, and you are using the right tool for the job. If you want recent conversation history, use get_conversation_messages instead of search_messages.
Conversations
| Tool | Use it to | Inputs |
|---|
search_conversations | Search channels and DMs by name or description. | q required. |
list_conversations | List accessible channels and DMs, optionally filtered by name or description. | Optional: q, limit (default 50, max 100). |
list_conversation_members | List active members in an accessible conversation, including humans, agents, and workspace roles. | conversation_id required. |
join_conversation | Join an accessible conversation as the current member. | conversation_id required. |
add_to_conversation | Add members to an accessible conversation. | conversation_id, member_ids. |
remove_from_conversation | Remove members from an accessible conversation. | conversation_id, member_ids. |
create_conversation | Create a new conversation. | name, member_ids. Optional: description, is_private. |
Messages
| Tool | Use it to | Inputs |
|---|
search_messages | Search accessible messages by keyword or semantic similarity. Rejects empty and wildcard-only queries. | q required. Optional: author, conversation, thread, after, before, mode (full-text or semantic). |
get_message | Fetch a single message with author, conversation, and content. | message_id required. |
get_conversation_messages | Browse recent messages from a conversation in reverse chronological order. | conversation_id required. Optional: author, limit, before. |
get_thread_replies | Fetch replies for a thread root message in chronological order. | message_id required. Optional: limit, after. |
send_message | Send a message to an accessible existing conversation. | conversation_id, markdown_content. |
reply_to_message | Reply in a message’s thread. If the target is already a reply, posts to the same thread. | message_id, markdown_content. |
react_to_message | React to a message with a supported standard emoji or active workspace custom emoji. | message_id, emoji. |
delete_message | Delete a message the current member can delete. | message_id required. |
Members
| Tool | Use it to | Inputs |
|---|
search_members | Search workspace members by name, email, or title. | q required. |
list_members | List active workspace members, optionally filtered by exact display name. | Optional: names. |
get_member | Fetch a single workspace member profile. | member_id required. |
send_direct_message | Reuse or create a DM and send a message. | member_ids, markdown_content. |
Calls / Jams
| Tool | Use it to | Inputs |
|---|
search_calls | Search call/Jam transcript content. | q required. Optional: conversation, after, before. |
list_calls | List accessible calls/Jams without a search query. | Optional: conversation, status (active or ended), recorded, after, before, limit (default/max 20). |
get_call | Fetch call/Jam metadata. | call_id required. |
get_call_transcript | Fetch a paginated transcript for a call. | call_id required. Optional: limit, cursor. |
Clipboard
| Tool | Use it to | Inputs |
|---|
search_clipboard | Search clipboard/context bundles saved from Ando. | q required. |
get_clipboard | Fetch a clipboard bundle and all saved items. | clipboard_id required. |
Workspace actions
| Tool | Use it to | Inputs |
|---|
invite_to_workspace | Invite a new member to the current workspace by email address. | email required. |
