Using the Console

This guide explains how to use Devify day to day: signing in, navigating the console, and running the Threadline workflow end to end.

Concepts

• Threadline — the core workflow that turns forwarded conversations (chat records, email threads, screenshots) into a structured summary.
• Chat — one processed conversation. Email in, structured summary out; you browse the results under Chats.
• Smart delivery (Relay) — an app that subscribes to workflow-completion events and delivers the result to an external target (Jira, Feishu Bitable).
• LLM config — a bound model provider entry. Devify uses a multimodal model for images and a text model for summarization.

Signing in

Open your Devify URL and sign in with username / password, or Google if OAuth is configured. New users can register at /register; password reset is available from the login screen.

Console overview

After signing in you land on the chats dashboard. Primary navigation:

Area	Path	What it's for
Chats	/chats	Browse processed conversations
TODOs	/todos	Action items extracted from conversations
App Center	/apps	Enable and open apps (e.g., Relay)
Settings	/settings	Profile and preferences
Billing	/billing	Credits and plan

Threadline workflow (end to end)

Forward by email  ──▶  Inbound (Haraka / IMAP)  ──▶  Worker processing
                                                        │  image understanding
                                                        │  LLM content analysis
                                                        │  summary + metadata
                                                        ▼
                                              Structured Chat  ──▶  Smart delivery
                                              (view under /chats)     (Jira / Feishu)

1. Send content in by email — forward the chat record, email thread, or screenshots to your Devify inbox address ({your-username}@{AUTO_ASSIGN_EMAIL_DOMAIN}). Both text and image attachments are supported.
2. Automatic processing — the worker runs multimodal image understanding and intent detection, LLM content analysis, then generates a summary and metadata (title, keywords/tags, TODOs). Status is visible on the chat while it runs.
3. Review the result — open Chats and select the conversation.
4. Deliver downstream (optional) — if a Relay channel is configured, the completed result is delivered automatically. See TODOs & smart delivery.

Working with chats

Under Chats each item is one processed conversation. Open one to see the generated summary (title and content), extracted metadata (keywords/tags), TODOs with any detected deadlines, and the original content and attachments.

Sharing — a chat can be shared via a read-only link for people without an account.