Operational AI for the investment lifecycle. Automate CIM analysis, DDQ completion & portfolio monitoring. Built for PE & private markets.
Users generally praise V7 Labs for its powerful AI-driven image annotation and workflow automation capabilities, highlighting efficiency and accuracy as main strengths. However, some users express concerns about its complex interface, which may require a steeper learning curve for new users. Sentiment around pricing is mixed, with some finding it reasonable given the feature set, while others suggest it could be more budget-friendly. Overall, V7 Labs holds a positive reputation among industry professionals, particularly for those involved in detailed AI and machine learning projects.
Mentions (30d)
0
Reviews
0
Platforms
2
Sentiment
0%
0 positive
Users generally praise V7 Labs for its powerful AI-driven image annotation and workflow automation capabilities, highlighting efficiency and accuracy as main strengths. However, some users express concerns about its complex interface, which may require a steeper learning curve for new users. Sentiment around pricing is mixed, with some finding it reasonable given the feature set, while others suggest it could be more budget-friendly. Overall, V7 Labs holds a positive reputation among industry professionals, particularly for those involved in detailed AI and machine learning projects.
Features
Use Cases
Industry
information technology & services
Employees
110
Funding Stage
Series A
Total Funding
$50.0M
My Claude.md file
This is my Claude.md file, it is the same information for Gemini.md as i use Claude Max and Gemini Ultra. # CLAUDE.md This file provides guidance to Claude Code (claude.ai/code) when working with code in this repository. ## Project Overview **Atlas UX** is a full-stack AI receptionist platform for trade businesses (plumbers, salons, HVAC). Lucy answers calls 24/7, books appointments, sends SMS confirmations, and notifies via Slack — for $99/mo. It runs as a web SPA and Electron desktop app, deployed on AWS Lightsail. The project is in Beta with built-in approval workflows and safety guardrails. ## Commands ### Frontend (root directory) ```bash npm run dev # Vite dev server at localhost:5173 npm run build # Production build to ./dist npm run preview # Preview production build npm run electron:dev # Run Electron desktop app npm run electron:build # Build Electron app ``` ### Backend (cd backend/) ```bash npm run dev # tsx watch mode (auto-recompile) npm run build # tsc compile to ./dist npm run start # Start Fastify server (port 8787) npm run worker:engine # Run AI orchestration loop npm run worker:email # Run email sender worker ``` ### Database ```bash docker-compose -f backend/docker-compose.yml up # Local PostgreSQL 16 npx prisma migrate dev # Run migrations npx prisma studio # DB GUI npx prisma db seed # Seed database ``` ### Knowledge Base ```bash cd backend && npm run kb:ingest-agents # Ingest agent docs cd backend && npm run kb:chunk-docs # Chunk KB documents ``` ## Architecture ### Directory Structure - `src/` — React 18 frontend (Vite + TypeScript + Tailwind CSS) - `components/` — Feature components (40+, often 10–70KB each) - `pages/` — Public-facing pages (Landing, Blog, Privacy, Terms, Store) - `lib/` — Client utilities (`api.ts`, `activeTenant.tsx` context) - `core/` — Client-side domain logic (agents, audit, exec, SGL) - `config/` — Email maps, AI personality config - `routes.ts` — All app routes (HashRouter-based) - `backend/src/` — Fastify 5 + TypeScript backend - `routes/` — 30+ route files, all mounted under `/v1` - `core/engine/` — Main AI orchestration engine - `plugins/` — Fastify plugins: `authPlugin`, `tenantPlugin`, `auditPlugin`, `csrfPlugin`, `tenantRateLimit` - `domain/` — Business domain logic (audit, content, ledger) - `services/` — Service layer (`elevenlabs.ts`, `credentialResolver.ts`, etc.) - `tools/` — Tool integrations (Outlook, Slack) - `workers/` — `engineLoop.ts` (ticks every 5s), `emailSender.ts` - `jobs/` — Database-backed job queue - `lib/encryption.ts` — AES-256-GCM encryption for stored credentials - `lib/webSearch.ts` — Multi-provider web search (You.com, Brave, Exa, Tavily, SerpAPI) with randomized rotation - `ai.ts` — AI provider setup (OpenAI, DeepSeek, OpenRouter, Cerebras) - `env.ts` — All environment variable definitions - `backend/prisma/` — Prisma schema (30KB+) and migrations - `electron/` — Electron main process and preload - `Agents/` — Agent configurations and policies - `policies/` — SGL.md (System Governance Language DSL), EXECUTION_CONSTITUTION.md - `workflows/` — Predefined workflow definitions ### Key Architectural Patterns **Multi-Tenancy:** Every DB table has a `tenant_id` FK. The backend's `tenantPlugin` extracts `x-tenant-id` from request headers. **Authentication:** JWT-based via `authPlugin.ts` (HS256, issuer/audience validated). Frontend sends token in Authorization header. Revoked tokens are checked against a `revokedToken` table (fail-closed). Expired revoked tokens are pruned daily. **CSRF Protection:** DB-backed synchronizer token pattern via `csrfPlugin.ts`. Tokens are issued on mutating responses, stored in `oauth_state` with 1-hour TTL, and validated on all state-changing requests. Webhook/callback endpoints are exempt (see `SKIP_PREFIXES` in the plugin). **Audit Trail:** All mutations must be logged to `audit_log` table via `auditPlugin`. Successful GETs and health/polling endpoints are skipped to reduce noise. On DB write failure, audit events fall back to stderr (never lost). Hash chain integrity (SOC 2 CC7.2) via `lib/auditChain.ts`. **Job System:** Async work is queued to the `jobs` DB table (statuses: queued → running → completed/failed). The engine loop picks up jobs periodically. **Engine Loop:** `workers/engineLoop.ts` is a separate Node process that ticks every `ENGINE_TICK_INTERVAL_MS` (default 5000ms). It handles the orchestration of autonomous agent actions. **AI Agents:** Named agents (Atlas=CEO, Binky=CRO, etc.) each have their own email accounts and role definitions. Agent behavior is governed by SGL policies. **Decisions/Approval Workflow:** High-risk actions (recurring charges, spend above `AUTO_SPEND_LIMIT_USD`, risk tier ≥ 2) require a `decision_memo` approval before execution. **Frontend Routing:** Uses `HashRouter` from React Router v7. All routes are defined in `src/routes.ts`. **Code Splitting:** Vite config splits chunks into `react-vendor`, `router`, `ui-vendor`, `charts`. **ElevenLabs Voice Agents:** Lucy's
View originalV7 Labs uses a subscription + contract + tiered pricing model. Visit their website for current pricing details.
Key features include: AI platform, purpose-built, for finance..
V7 Labs is commonly used for: Automating document processing workflows, Enhancing data labeling efficiency for machine learning projects, Streamlining compliance checks in regulated industries, Facilitating real-time collaboration on data-driven projects, Improving accuracy in data extraction from complex documents, Accelerating the training of AI models with expert-built agents.
V7 Labs integrates with: Slack, Google Drive, Microsoft Teams, Jira, Trello, Salesforce, Zapier, Asana, Dropbox, Box.
My First 6 Months at an AI Startup | Meet Sophie, BDR
Mar 12, 2026
