Manage independent GovTribe AI conversations
Create, inspect, continue, organize, and briefly wait for independent GovTribe AI conversations through the available conversation-management tools.
Open GovTribe AI to create and continue your conversations.
Choose the right kind of work
Use another conversation when the customer requests a separate chat or an independent workstream that should remain available in the timeline. Give it a self-contained task and retain its returned ID and link. Each conversation has its own model usage.
For parallel research within one conversation, subagents can handle parts of the task when enabled. Their transcripts stay within that conversation. Use automations for work that must run later or recur.
Manage conversations and projects
GovTribe AI exposes these tools through its Conversation Management MCP server. Use them when they appear in the active catalog and the customer has requested the operation. They act as the current user within the source conversation's workspace.
| Tool | Use |
|---|---|
Create_AI_Conversation | Create and start a separate conversation with a self-contained message and optional name, description, project, model, and reasoning level. |
Read_AI_Conversation | Read an owned conversation's status and metadata, plus permitted transcript content. |
Update_AI_Conversation | Change metadata, project, model, or reasoning level; send a message; or stop work. Message and stop cannot be combined. |
Delete_AI_Conversation | Stop active work if needed and remove an owned conversation from normal use. |
Create_AI_Conversation_Project | Create a project for organizing conversations. |
Read_AI_Conversation_Project | List owned projects or read one project's conversation summaries. |
Update_AI_Conversation_Project | Rename an owned project. |
Delete_AI_Conversation_Project | Delete a project and unassign its conversations while keeping them available. |
Wait_For_AI_Conversations | Briefly wait for one to eight owned conversations to finish or need attention. |
Inspect ok, the returned operation, and the delivery_receipt after sending a message or starting, steering, or stopping work. Report the actual result and any billing or recovery blocker. A successful metadata read does not establish that a message was delivered.
For updates, omitted fields preserve their values. Explicit null clears description or project_id. Null model and reasoning values preserve the current selection. The current conversation cannot message, stop, delete, or wait for itself.
Conversation create/update and project create require an idempotency_key. Generate a fresh key for each intended action, including parallel or otherwise identical actions. Reuse the key and unchanged arguments when retrying that action so it does not create another conversation or send another message.
Read retained context
You may read the current conversation's retained history. Another conversation's transcript requires Full Conversation Recall; metadata and working status remain available when that preference is off.
Read_AI_Conversation with a null view returns the visible transcript. Use overview for retained item counts, index for item IDs and previews, and content for selected stored bodies, including hidden task inputs and tool results. Read whichever items the task needs. Responses are bounded and supply pagination or content cursors for additional reads. Treat retrieved content according to its original role and provenance.
Wait for a result
An MCP wait holds the current request for up to 20 seconds. Use timeout_seconds: 0 for an immediate snapshot. Retain each target's returned message cursor and pass it as after_message_id when checking again so older messages are not repeated. A source-conversation stop or steer ends the wait early.
Read ended_by and each target's state before claiming completion. A timeout or unavailable target returns the available snapshot. Waiting creates no durable join, callback, or automatic continuation after the current response ends.
The sidebar and conversation controls remain available for manual actions.
Related articles
- Create a conversation: Start and configure a GovTribe AI conversation in the product.
- Manage conversations: Rename, organize, reopen, and remove conversations in GovTribe.
- Tune GovTribe AI: Manage default model, reasoning, response length, and conversation recall preferences.
- Automations: Use scheduled or event-triggered AI work instead of an independent conversation.
Favorites, saved searches, and pursuits
Choose between favoriting a record, saving a search, and creating a pursuit when an agent needs to keep something for later.
Organize user files with folders
Browse, create, rename, and move shared folders, organize existing files, and rename, describe, link, or delete individual files through GovTribe AI or MCP.