Layi Docs
Threads, Messaging, and Voice
Keep every task, attachment, and call in one durable channel
Why threads matter
Threads are the canonical record for every conversation between your team and Layi’s workforce. Each thread captures prompts, attachments, voice call controls, and a timestamped activity log so compliance and operators always know what happened and why.
Starting a thread
From the dashboard, click New Task or open an existing project and press Start Thread. Pick the agent, choose the delivery surface (chat or phone), and describe what you need. Threads inherit the project’s permissions so only authorized teammates can view or send messages.
- Choose agent + project
Switch agents without leaving the composer; thread metadata keeps the association.
- Set urgency
Optional priority tags flow into the activity log and downstream notifications.
- Save drafts
Drafts auto-save per user, letting you assemble long instructions before running.
Messaging & file attachments
The composer supports rich text, snippets, and unlimited attachments. Files are stored in your MinIO bucket, encrypted, and linked to the thread timeline with preview + download controls.
- Large files
Files over 200 MB upload via a resumable uploader so network drops never corrupt the attachment.
- Retention
Attachments inherit your thread retention configuration; purge schedules are visible in the activity log.
Voice & call controls
Voice sessions run inside the thread so operators can jump in mid-call. The control bar exposes real-time mute, unmute, hold, and end-call actions, and every toggle is logged for auditing.
- Mute / Unmute
Temporarily pause the agent’s microphone when sharing sensitive details.
- Hold
Place the caller on hold while a human reviews the transcript or fetches approvals. Music/voice prompts are customizable per agent.
- End
Terminate immediately; the call summary and final transcript appear in the timeline seconds later.
Activity log dashboard
Every thread exposes a side-panel activity log summarizing runs, tool usage, clarifications, and human interventions. Use it to troubleshoot or export evidence for governance reviews.
- Filters
Slice by agent, date range, or outcome (clarification requested, completed, aborted).
- Exports
Download JSON or CSV snapshots for audits; exports contain the same IDs used in the API.