Scriben is an AI-powered smart pen that captures real-world conversations and transforms them into structured, actionable work across a user's existing software stack. It combines a physical pen with an intelligent agent that remembers context, prepares follow-ups, and executes approval-gated actions in tools like email, calendars, and CRMs. The device records locally for up to 18 hours and syncs only when the user chooses, ensuring privacy and control over all captured data.
Funding
Funding not disclosed
Founders
Product
Problem
Real-world conversations—in meetings, on sales calls, or in creative sessions—contain critical decisions, commitments, and context that are often lost once the moment passes. Traditional note-taking distracts from the conversation, and existing AI tools require screens and cannot capture the full nuance of in-person interactions, leaving knowledge workers to manually reconstruct details and follow up on action items.
Solution
Scriben provides a hardware AI pen that captures conversations in person and transforms them into a private intelligence layer. The pen records audio locally for up to 18 hours without needing a device, then syncs to a workspace where speech is converted into structured knowledge—identifying who said what, decisions made, and open tasks. From this memory, Scriben's agent prepares follow-up emails, calendar events, CRM updates, and documents, all gated behind user approval before anything is sent. The system connects natively to 9,000+ apps like Gmail, Slack, and HubSpot, and can also provide permissioned memory access to AI assistants like ChatGPT and Claude via MCP, ensuring work moves forward without leaving the user's control.
Target Audience
Primary customers are knowledge workers, sales professionals, healthcare providers, and creators who spend significant time in meetings and need to capture detailed context without disrupting their workflow, including those in client-facing roles and clinical settings.
Features
- Hardware AI pen with local audio capture for up to 18 hours, requiring no phone or other device during recording
- Automatic speech-to-text transcription with speaker identification and extraction of decisions, commitments, and action items
- Approval-gated agentic workflows that draft and send emails, book calendar events, update CRM records, and create files only after user confirmation
- Native integrations with Gmail, Outlook, Calendar, Drive, Teams, Slack, Linear, Jira, HubSpot, and Notion, plus 9,000+ apps via API
- MCP (Model Context Protocol) support for granting permissioned memory access to AI assistants like ChatGPT, Claude, and Zoom AI Companion
- Privacy-first architecture with offline capture, user-controlled sync, no model training on customer data, and HIPAA-compliant BAA-backed onboarding for healthcare use