Open Source · AI-Powered Command Center

Stop Reacting.
Start Commanding. The open-source, local-first AI notification command center for Slack, Gmail, GitHub, Jira, Notion, Outlook, Calendar & more.

Laya enriches your daily workflows with AI — whether you connect one platform or many. It researches context, drafts responses, and delivers ready-to-approve action cards before you even open the app — using local LLMs via Ollama and LM Studio, or your own API keys.

Download Laya
See How It Works

All platforms · latest release · View source on GitHub

Open source. Local-first. Works with one platform or many. Your API keys, your data.

Laya - Your AI command centre | Product Hunt Laya on PeerPush Laya — #1 ON THE CAPITAL

What is Laya?

Laya is an open-source desktop app that brings your work tools — Gmail, Google Calendar, Slack, Jira, GitHub, and more — into one intelligent feed. When a new email, meeting invite, ticket, or pull request arrives, Laya's AI reads it, gathers related context from your connected tools, and turns it into an action card with a suggested response — drafted and ready for your review.

You stay in control: nothing is sent or changed on your behalf without your explicit approval. And because Laya is a desktop-only app with no servers behind it, your emails, events, and messages are stored entirely on your computer — and when you use a local model, they never leave your machine.

Laya Card View — grouped action cards in 3-column grid with intelligence reports
9+
Platform Integrations
6
AI Personas
10-Stage
Intelligence Pipeline
100%
Open Source & Local-First

Your workflow is broken.

Whether you juggle eight tools or just an overflowing inbox — every notification demands attention and none of them come with context.

Without Laya

  • Open Slack notification. No context. Search Jira for the ticket. Open Bitbucket for the PR. Piece together what happened. 15 minutes gone.
  • Miss a critical email buried under 40 promotional messages you need to manually triage.
  • Start your morning with 30 minutes of "catching up" — reading threads, checking calendars, reviewing PRs.
  • Context-switch 50+ times per day between tools. Each switch costs 23 minutes of focus.

With Laya

  • Open Laya. AI already researched the context, cross-referenced the ticket, found related PRs, and drafted a response. One click to approve.
  • AI intelligence reports flag the recruitment email as legitimate, auto-archives 39 promotional ones, surfaces what matters.
  • Daily briefing synthesizes overnight activity into Events, Action Items, and Key Updates — ready at 7 AM.
  • One unified feed. Every platform. Every notification pre-processed with intelligence. Zero context switching.

Intelligence, not notifications.

Every card Laya delivers has already been researched, classified, and prepared for action.

Intelligence Reports

Every notification gets an AI-generated intelligence report. Sender reputation, cross-platform context, pattern recognition from your history, and actionable insights — all before you read a single word.

Pre-Drafted Actions

Draft replies, code fixes, briefing summaries — ready for one-click approval. The AI does the work. You just say "go" or refine the details.

Daily Briefings

Wake up to an AI-synthesized morning briefing. Events & meetings, action items, key updates — organized by priority across all your spaces.

Coding Agents

Run coding agents (Claude Code, Gemini CLI, Codex, Pi CLI) directly from any card. Watch them work in real-time and guide their actions from the workspace.

Smart Classification

Every event is automatically categorized by persona (Engineer, Comms, Ops, Finance, HR, Sales), priority (Critical to Low), and category — using a multi-model AI pipeline that learns from your corrections.

Spaces

Organize your life into separate contexts — Work, Personal, side projects. Each space has its own sources, model configs, and API keys.

Coherence Search

Search for any entity across all platforms and trace its full lifecycle — from Jira ticket to PR to Slack thread. Hybrid retrieval blends semantic vectors with keyword (BM25) ranking, and AI-generated narratives tell the complete story.

Cross-Platform Composer

Execute outbound actions from a single interface — send emails, post Slack messages, comment on tickets, and more. AI-assisted drafting with full preview before sending.

Bookmarks

Bookmark important cards for quick access. Filter your feed to show only bookmarked items — no matter the date or status.

Omni

A rolling cross-platform summary that answers "where am I right now?" at a glance. Four temporal layers — from items needing attention to long-term milestones — compressed by AI.

Context Association

Automatically links related cards across platforms using semantic similarity and LLM confirmation. Learns editable grouping rules from your link/unlink corrections to improve accuracy over time.

Agent Studio

Launch a fully configured agent session from any card. Choose your runtime, set an execution mode, configure directories, and write detailed prompts — with Plan or Accept Edits modes for full control.

Budget Tracking

Monitor LLM costs broken down by feature (Pulse, Omni, Chat, Coherence) and pipeline step. Set monthly spending caps with automatic pause when limits are reached.

Run on Your AI Subscription

Point any pipeline stage at an installed CLI agent — Claude Code, Codex, Gemini, or Pi — and Laya runs on that agent's own quota instead of a separate API key. A window-based usage budget auto-pauses ingestion before you hit a limit and resumes when the window resets.

Rules & Automation

Filter noise, auto-classify, and fire automated actions — tag, route, run an agent, send a reply — with optional AI-evaluated conditions. Every firing lands in a searchable log so you can see exactly what ran and why. Edit any rule straight from chat.

Cloud or Self-Hosted

Works with cloud SaaS and on-prem instances alike — point a repo at Bitbucket Server / Data Center or GitHub Enterprise via a per-repo host, and Laya routes ingestion and outbound actions to the right place.

How Laya works

A multi-stage AI pipeline transforms raw notifications into ready-to-approve intelligence.

Ingest

Events flow in from Gmail, Slack, Jira, GitHub, Calendar — connect as many or as few as you need.

Route & Classify

Fast AI classifies priority, category, persona, and builds a research plan.

Research & Draft

Specialized workers research context, draft replies, spawn coding agents, gather intelligence.

Deliver

Polished action cards land in your feed with intelligence reports and one-click actions.

Intelligence

Every card comes pre-researched

Laya doesn't just forward notifications. It investigates them. Each card includes an AI intelligence report with cross-platform context and pattern-matched history.

  • Sender reputation and domain verification
  • Cross-referenced past events from memory
  • Tracking link analysis and safety flags
  • Confidence scoring on every classification
Laya intelligence report showing sender analysis and cross-platform context
Views

Your feed, your way

Switch between list view for rapid triage or card view for visual scanning. Sort by newest, priority, category, or platform. Context-grouped cards cluster related items automatically.

  • List and Card view modes with context grouping
  • Filter by status, priority, category, space
  • Bulk actions — approve, dismiss, archive
Laya list view showing multiple view options
Briefings

Your morning, already organized

Every morning, Laya synthesizes overnight activity across all your platforms into a structured briefing. Events, action items, and key updates — prioritized and ready.

  • Automated daily synthesis at your preferred time
  • Events & meetings with context
  • Action items ranked by priority
  • Key updates filtered from noise
Laya daily briefing showing events, action items, and key updates
Summary

Your day at a glance

The Day Summary distills your entire day into a single structured view. Action items with pending counts, key updates filtered from noise, and space-level breakdowns — all generated and updated automatically.

  • Action items with pending status and priority
  • Key updates from all connected platforms
  • Space-level breakdown for multi-context workflows
Laya Day Summary — structured view of action items and key updates
Workspace

A full workspace for every agent session

Run coding agents — Claude Code, Gemini CLI, or Codex — directly from any card in your feed. Each session gets a dedicated workspace with a live timeline, context panel, and generated files.

  • Run an agent from any card with a single click
  • Live timeline of agent actions, tool calls, and progress
  • Full context panel with card info, implementation plan, and related entities
  • Interactive messaging — guide, question, or redirect mid-session
Laya agent workspace — full workspace with timeline, context panel, and agent output
Chat

Ask anything about your data

Laya's built-in chat uses hybrid semantic + keyword search across all your ingested events. Ask natural language questions and get instant, context-rich answers — with links back to the original cards and events.

  • Hybrid semantic + keyword (BM25) search across all platforms and events
  • Natural language queries — no filters needed
  • Direct links to related cards and action items
  • RAG-powered retrieval from your local data
Laya chat with semantic search — ask natural language questions about your data
Omni

Your big picture, always current

Omni is a rolling cross-platform summary that synthesizes all your professional activity into a single unified view. It answers "where am I right now?" — especially powerful after time away or for sprint retrospectives.

  • Four temporal layers: Attention, Recent, Period, and Milestone
  • Progressive summarization — older layers compress automatically
  • Pin important items so they survive compression
  • Time-travel slider to view previous snapshots
Laya Omni — rolling cross-platform summary with progressive summarization
Coherence

Trace any entity across every platform

Coherence is Laya's cross-platform entity search. Search for a person, ticket, PR, or project and see its complete lifecycle — every event, every platform, one chronological story.

  • Semantic + fuzzy + entity search merged via Reciprocal Rank Fusion
  • Auto-clustered by entity relationships
  • AI-generated narrative with chapter detection
  • Export traces as markdown, save and rerun later
Laya Coherence — trace entities across platforms with AI-generated narratives
Context Association

Related cards, automatically linked

Laya's context association engine uses semantic similarity and LLM confirmation to automatically group cards that belong to the same real-world context — even across different platforms and entity types. Link or unlink groups manually, and Laya learns from every correction.

  • Semantic matching with confidence thresholds (auto-confirm + LLM-confirm)
  • Manual merge and unlink via Link Dialog
  • Learning pipeline extracts grouping rules from corrections every 6 hours
  • Respects user-split decisions — never re-merges what you separated
Laya Context Association — related cards automatically grouped across platforms
Inline Agents

Launch an agent from any card

Every card in your feed has a Run Agent action. Provide an optional focus prompt and start the agent instantly — no context switching, no setup. The card's intelligence report and related entities are automatically injected as context.

  • One-click agent trigger from any card detail view
  • Optional focus prompt to guide the agent's investigation
  • Card context and related entities injected automatically
  • Live progress streaming via workspace timeline
Laya Run Agent — launch a coding agent directly from card detail view
Agent Studio

Full control over your agent sessions

For complex tasks, the Agent Studio gives you full control. Choose your agent runtime, select an execution mode, configure working directories, attach reference files, and write detailed prompts — all before the agent starts.

  • Choose from Claude Code, Gemini CLI, or Codex
  • Plan mode for review-first workflows, or Accept Edits for autonomous execution
  • Configure working directories and attach reference files
  • Auto-provisioned research workspaces for investigation tasks
Laya Agent Studio — configure agent runtime, mode, and prompt before launching
Composer

One interface for every outbound action

The Composer lets you execute actions across all connected platforms from a single modal. Send emails, post Slack messages, comment on Jira tickets, create GitHub issues, and more — with AI-assisted drafting and full preview before sending.

  • 9 platform backends: Gmail, Outlook, Jira, Notion, GitHub, Bitbucket, Linear, Calendar, and more
  • AI Assist generates context-aware drafts from your card data
  • Per-platform action types — compose email, post comment, create ticket
  • Multi-account support with connection health monitoring
Laya Composer — unified outbound action interface across all connected platforms
Custom Prompts

Every stage is yours to shape

Laya's pipeline runs through distinct LLM stages — router, stager, the six persona workers, briefings, summaries, semantic chat, and more. Drop a markdown file into ~/.laya/prompts/ for any stage and Laya uses your prompt instead of the default. No fork, no rebuild.

  • Override 20+ pipeline stages individually — change one, keep the rest as default
  • Tune persona voices, classification rules, or briefing style to your team
  • Delete the file to revert — defaults always come back
  • Reload picked up on engine restart — version-control your prompts alongside your dotfiles
~/.laya/prompts/
├─ router.md          # classification
├─ stager.md          # action drafting
├─ engineer.md        # persona voice
├─ comms.md
├─ briefing.md        # daily digest
├─ omni.md            # rolling summary
└─ … 16 more
MCP Server

Bring Laya into any AI harness

Whenever Laya is running, a Model Context Protocol server is reachable at http://127.0.0.1:8420/mcp/. Grab a token from Settings → MCP and wire any MCP-compatible client — Claude Desktop, Cursor, VS Code, custom agents — straight into your live data.

  • Three scope toggles — Read, Write, Egress — gate what clients can call
  • Bearer-token auth or loopback-only mode, your call
  • Full surface: search cards, semantic search, mutate, and trigger outbound actions across all 9 connected platforms
  • Same transport powers Laya's in-app coding agents — one path, one policy
{
  "mcpServers": {
    "laya": {
      "command": "npx",
      "args": [
        "mcp-remote",
        "http://127.0.0.1:8420/mcp/",
        "--allow-http",
        "--header",
        "Authorization: Bearer lyat_…"
      ]
    }
  }
}

Six AI personas.
Zero context switching.

Laya automatically routes each notification to the right specialist persona, so you get expert-level output every time.

Engineer

Spawns coding agents, analyzes PRs, reviews build failures, drafts code fixes, and surfaces implementation plans with full repo context.

Comms

Drafts email replies, Slack responses, and PR comments. Analyzes sender intent, checks communication history, and prepares context-aware responses.

Ops

Aggregates calendar events, manages scheduling conflicts, synthesizes meeting prep, and monitors operational updates across all platforms.

Finance

Processes invoices, expense reports, and budget updates. Flags financial anomalies and prepares cost summaries across your accounts.

HR

Handles people-related events — onboarding, leave requests, team announcements, and HR policy updates. Recognizes internal people operations.

Sales

Tracks prospects, deals, and pipeline activity. Summarizes customer interactions and flags deal-stage changes that need attention.

Connects to everything you already use

Connect one tool or all of them. Laya normalizes everything into a unified, AI-enriched event stream.

Gmail
Slack
Jira
GitHub
Bitbucket
Calendar
Outlook
Linear
Notion
More via n8n

Your data stays yours

Laya is local-first by design. Your data never leaves your machine unless you choose a cloud model.

Local-First Architecture

SQLite database, ChromaDB vectors, and all configs stored on your machine. No cloud account required.

Your Keys, Your Models

Bring your own API keys. Use Claude, GPT, Gemini, or run Ollama locally for complete air-gapped privacy.

3-Tier Privacy Classification

Sensitive content (DMs, emails) can be routed to local models only. Metadata stays separate from content.

Ready to take
command?

Laya is free and open source. Connect your first tool, run the engine, and let AI handle the rest.

Download Laya
Read the Docs

All platforms · latest release · View source on GitHub