Day 8 of #100DaysOfAI - Maps how you actually work. Chrome Extension + VS Code Extension.
FlowMap AI silently observes your workflow across SaaS tools (Chrome) and your codebase (VS Code), then:
- Detects every tool transition, context switch, and navigation pattern
- Maps your actual workflow as a directed graph with frequency-weighted edges
- Finds bottlenecks, central dependencies, and time sinks
- Recommends specific automations to save hours per week
No surveys. No self-reporting. No guesswork. Pure behavioral data.
Tracks how you move between Slack, Jira, GitHub, Gmail, Figma, and 60+ other tools. Discovers patterns like:
- "Every time I get a Slack notification, I open Jira, then switch to GitHub" (Bug Fix Cycle)
- "I check Gmail → Salesforce → Google Docs 8x/day" (Sales Workflow)
- "I context-switch between 12 tools per hour" (Focus problem)
Tracks file navigation, saves, terminal usage, and debug sessions inside your IDE. Discovers:
- Which files you touch together (coupling signals)
- Hot files that get edited constantly (refactor candidates)
- Your actual development flow: source → test → config → deploy
┌─────────────────────────────────────────────────────────────────┐
│ FlowMap AI Pipeline │
│ │
│ ┌──────────────┐ ┌──────────────┐ ┌──────────────┐ │
│ │ PATTERN │──▶│ FLOW │──▶│ PROCESS │ │
│ │ DETECTOR │ │ MAPPER │ │ ANALYZER │ │
│ │ (Scout) │ │ (Analyst) │ │ (Executor) │ │
│ └──────────────┘ └──────────────┘ └──────────────┘ │
│ │
│ 200+ SaaS tools Directed graph Bottleneck detection │
│ N-gram sequences Mermaid flowcharts Automation opportunities │
│ Time patterns Edge weights Efficiency scoring │
│ Context switches Process matching Time savings calc │
│ Tool classification Bottleneck detect Recommendations │
└─────────────────────────────────────────────────────────────────┘
- 200+ SaaS tool classifier — recognizes Jira, Slack, GitHub, Figma, etc. from URLs
- N-gram sequence detection (finds repeated 2-5 step workflows)
- Time pattern analysis (peak hours, peak days, per-tool activity)
- Context switch measurement with focus scoring
- Tool usage stats with time-per-tool estimation
- Builds directed transition graph from raw events
- Edge weights = transition frequency
- Bottleneck detection: central dependencies, work sinks, time sinks, slow transitions
- Generates Mermaid flowcharts for visualization
- Process classification: matches against 8 known patterns (Bug Fix Cycle, Feature Dev, Incident Response, etc.)
- Automation opportunity finder
- Actionable recommendation engine
- Time savings calculator (weekly/monthly/yearly + dollar value)
- Optional LLM synthesis via Groq for executive summaries
- Workflow efficiency scoring (0-100, A+ to F)
| Feature | FlowMap AI | Survey/Manual Process Mapping |
|---|---|---|
| Data collection | Automatic (zero effort) | Manual surveys, interviews |
| Accuracy | Behavioral truth | Self-reported (biased) |
| Works without LLM | ✅ Full pipeline | N/A |
| Real-time | Live tracking | Quarterly review |
| Granularity | Every click/transition | High-level guesses |
| Cost to deploy | Install extension | Hire consultant ($50K+) |
| Task | Consultant | FlowMap AI |
|---|---|---|
| Map team's actual workflow | 2-4 weeks | Instant (after 1 day tracking) |
| Find bottlenecks | Interviews + analysis | Automatic |
| Recommend automations | Subjective | Data-driven |
| Track changes over time | New engagement | Continuous |
| Total cost | $50K-$200K | Free / $49/mo |
# 1. Clone repo
git clone https://github.com/sathishlella/flowmap-ai.git
cd flowmap-ai/chrome-extension
# 2. Load in Chrome
# Go to chrome://extensions → Enable Developer Mode → Load Unpacked → Select chrome-extension folder
# 3. Click the FlowMap AI icon → Start Tracking# 1. Navigate to VS Code extension
cd flowmap-ai/vscode-extension
# 2. Package
npx @vscode/vsce package
# 3. Install
code --install-extension flowmap-ai-1.0.0.vsix
# 4. Start tracking
# Cmd+Shift+P → "FlowMap AI: Start Tracking"| Feature | Details |
|---|---|
| Tool Detection | 200+ SaaS tools auto-classified |
| Workflow Patterns | N-gram detection finds repeated sequences |
| Context Switches | Measures rapid switching between tool categories |
| Focus Score | 0-100 score based on switching frequency |
| Efficiency Grade | A+ to F based on bottlenecks + tool sprawl |
| Flow Visualization | Mermaid graph of tool transitions |
| Privacy | All data stays in local chrome.storage (never leaves browser) |
| Feature | Details |
|---|---|
| File Tracking | Classifies files: UI, API, test, config, docs, utility |
| Save Tracking | Knows which files you edit most |
| Terminal Tracking | Records terminal open events |
| Debug Tracking | Start/stop debug session awareness |
| Hot Files | Detects files that dominate your workflow |
| Dashboard | Webview panel with full analytics |
| Persistence | Events survive VS Code restarts |
| Tier | Price | Target |
|---|---|---|
| Free | $0 | Individual developers |
| Team | $12/user/mo | Engineering teams (5-50) |
| Enterprise | $29/user/mo | Company-wide process discovery |
| Platform | Custom | Process mining firms, consultancies |
- Engineering Leaders: "See how your team actually works vs. how you think they work"
- Process Consultants: "Replace 4-week engagement with 1-day automated discovery"
- IT Departments: "Find Shadow IT — which unsanctioned tools are your teams using?"
- HR/People Ops: "Measure tool adoption after new tool rollout"
Process Mining market: $5.4B (2025), growing 40% YoY. Comparable companies: Celonis ($13B valuation), UiPath ($7B), ProcessGold (acquired by UiPath).
- Chrome: All data in
chrome.storage.local— never leaves the browser - VS Code: All data in
globalState— never leaves the machine - No tracking server — zero telemetry, zero cloud dependency
- No API keys needed — core engine is 100% offline
- Enterprise tier adds optional centralized dashboard (self-hosted)
- Core Engine: Vanilla JavaScript — zero dependencies, 100% portable
- Chrome Extension: Manifest V3, Service Worker, chrome.storage API
- VS Code Extension: VS Code Extension API, Webview panels
- LLM: Groq (optional) — executive summaries only
- Architecture: 3-agent pipeline (Pattern Detector → Flow Mapper → Process Analyzer)
MIT — Built by Sathish Lella as Day 8 of #100DaysOfAI