Open-Source Laya Turns Slack and Gmail Chaos Into Approved Actions
An open-source desktop app funnels Slack, Gmail, GitHub, Jira and Notion pings through local LLMs and turns each into an approve-or-dismiss Action Card.
- Laya is an open-source local-first notification command center for Slack, Gmail, GitHub, Jira and more.
- Uses local LLMs via Ollama or LM Studio, with BYOK for Claude, GPT, Gemini through LiteLLM.
- Routes events to persona-specific agents (Engineer, Comms, Ops, Sales, HR, Finance) that pre-stage Action Cards.
- Hybrid retrieval combines ChromaDB vectors with SQLite FTS5 BM25 via Reciprocal Rank Fusion.
- Ships as Tauri desktop app with Python FastAPI engine and local n8n integration gateway.
- Includes per-feature budget caps, auto-pause on quota, and OS keychain for API keys.
Laya turns work notifications into prewritten actions
Laya is an Apache-2.0-licensed desktop app that gathers events from Slack, Gmail, GitHub, Jira, Notion, Outlook, and Calendar. An AI agent classifies each event, retrieves related context, and prepares a reply or other action for approval.
The repository has 6 GitHub stars. Its four releases indicate an early project, but the activity shows developer interest in running notification triage on a local desktop stack.
From incoming event to Action Card
Each incoming event becomes an Action Card containing its priority, supporting context, and a proposed response. Depending on the source, that response might be an email draft, a pull-request comment, a ticket update, or a request to dismiss the notification.
Approval sends the staged action back to the originating service. Users can also edit or reject a card, preserving a human checkpoint before Laya writes to connected tools.
Four stages separate reading from writing
- n8n, running locally on Node.js, receives or fetches events from connected services.
- The workflow normalizes each event and passes it to Laya’s Python engine.
- The engine classifies the event, retrieves related information, and creates an Action Card for the Tauri and Svelte desktop interface.
- After approval, n8n executes the requested action through the source service.
| Component | Implementation | Responsibility |
|---|---|---|
| Workflow layer | n8n and Node.js | Connects services, normalizes events, and executes approved actions |
| Agent engine | Python 3.10+, FastAPI, and asyncio | Routes, researches, prioritizes, and stages work |
This story is for Pro members
You've reached the end of the free preview. Upgrade to AlphaSignal Pro to read the full article - and everything else behind the paywall.
Comments
