Margot v3.0

A model-selectable AI assistant for macOS and iOS with 84 Google Cloud tools, scheduled automation, and memory that persists.

Choose your coordinator from any model on OpenRouter, with ten of the latest models already configured. Orchestrate tasks across Google Cloud, Drive, Gmail, & local files, automate browser tasks, and create scheduled workflows, backed by four-layer cognitive memory and context laundering that keeps every conversation sharp and affordable.

0:00 / 0:00

Capabilities

Google Cloud

84 tools across Analytics 4, Search Console, Merchant Center, Gmail, Drive, Sheets, Calendar, and PageSpeed Insights. Multi-account OAuth with automatic routing by connection type.

Research

Three tiers: standard chat (5–15s), light research with subtopic pipelines (30–90s), and deep research with 4 independent agents and cross-pollination (90–180s).

Desktop & Browser

21 desktop tools with three-level shell safety. Browser automation via Playwright with 48 skills, deterministic recovery, and release gating.

Media Generation

Image generation via Gemini 3 Pro (up to 4K), video via Veo 3.1 and Sora 2, voice synthesis via ElevenLabs with streaming.

Scheduled Automation

Run full coordinator workflows unattended on cron, interval, or slot schedules. Verification contracts, mutation guardrails, and cooperative cancellation for safe background execution.

Context Laundering

Large tool payloads are processed before LLM reinjection—raw, summarized, or truncated based on content type. Integrated across chat, research, and scheduled task runtimes.

Generated Skills

Create new tools from natural language. The skill generator produces Python, validates via AST analysis and automated tests, then hot-registers into the coordinator’s tool set—no restart required.

Architecture

A model-selectable coordinator runs a ReAct loop (up to 30 iterations) and delegates to specialized subagents. Route through any model on OpenRouter—with automatic resilience that retries and falls back on pre-first-chunk failures.

Default Models

Grok 4.1 Fast Claude Sonnet 4.6 Gemini 3.1 Pro GPT-5.4 GLM-5 MiniMax M2.5 MiniMax M2.7 Kimi K2.5 MiMo V2 Pro MiMo Omni
Research 4× Grok 4.1 Fast, cross-pollination between rounds, 90–180s
Browser Gemini 3 Flash loops, 48 skills, deterministic recovery
Tauri 2.0 React FastAPI PostgreSQL + pgvector Tailscale SwiftUI

Memory

Four-layer cognitive memory inspired by Mem0. Memories are extracted after each conversation and assembled in under 150ms.

01

Short-Term

Last 20 messages from the current conversation, kept in PostgreSQL.

02

Semantic

Long-term facts and preferences stored as vectors with HNSW indexing and automatic deduplication.

03

Episodic

Past conversation summaries with importance scoring—top 3 retrieved per query.

04

Procedural

Reusable workflow patterns detected from tool usage sequences with success tracking.

iOS App

A native SwiftUI client with the same capabilities as the desktop. Connect securely over Tailscale from anywhere.

Skills Library

Browse and manage all 48 skills from your phone.

Scheduled Tasks

Create, monitor, and review automation runs on the go.

Tool Traces

Inspect every tool call with inline reasoning and full execution traces.

Model Selection

Pick your coordinator model and tune settings per conversation.

Internet Provider

Switch between Brave and Firecrawl search providers per chat.

Push Notifications

APNs-backed alerts for scheduled task outcomes and reminders.

Get Margot

See what Margot can do for your workflow.

~$0.001/query standard ~$0.01 light research ~$0.016 deep research
Request a Demo

macOS 13+ · iOS 17+