Lutra AI

Lutra AI develops an AI agent that converts natural-language instructions into executable code, generating and running automated workflows that connect a user's existing applications. The agent first converses with the user to clarify goals, then writes code that implements the workflow; because the workflows are code, they can be inspected, executed repeatedly, and scheduled to run unattended. Typical use cases described by the company include contact and CRM data enrichment, extracting structured data from PDFs and websites, web and internet research, email management and personalized outreach, expense reporting, and converting data into reports and visualizations.

The platform ships with built-in AI tools for data processing (structured extraction, classification, document drafting and conversion, template rendering, visualization) plus connectors to a broad set of SaaS applications, including Google Workspace (Sheets, Docs, Gmail, Calendar), Microsoft Outlook and OneDrive, Slack, GitHub and GitHub Actions, HubSpot, Airtable, Notion, Stripe, Xero, Apollo, Affinity CRM, ActiveCampaign, Maximizer CRM, AppFolio, DocSend, Harvest, Keepa, LinkedIn posting, weather data, and a Y Combinator company dataset. Custom integrations can be created via MCP (Model Context Protocol), REST/HTTP APIs, OpenAPI specifications and databases such as Supabase, and the company lists a further set of integrations as forthcoming (including Salesforce, Zapier, Microsoft Excel, QuickBooks Online, Microsoft Teams and others). Recurring tasks can be saved as "Playbooks" — reusable workflows that can be scheduled and shared, with organization-wide sharing on enterprise plans.

The company's website also promotes MintMCP by Lutra, positioned as a way to securely connect enterprise data and applications to AI. Lutra states it is SOC-2 compliant, uses OAuth for integrations, gives users visibility into all actions and generated code, and does not retain customer data long-term unless requested.

Founding story

Jiquan Ngiam, who worked at Coursera and later Google, repeatedly encountered engineering functions that could have been automated for non-developers but lacked the resources to do so. In early 2023 he gathered five friends, including former Google colleague Vijay Vasudevan, to examine how AI models capable of reasoning, planning and coding could interconnect everyday software reliably and securely for non-engineers. The startup began in April 2023 and focused early experiments on mundane tasks such as taming a busy inbox, managing Slack connections with customers, and internet research. Sources differ on founder attribution: TechCrunch and Starter Story credit Jiquan Ngiam as co-founder and CEO (with Vijay Vasudevan among the founding group), while one aggregator profile names Bo Zhi See and Jiquan Ngiam as the 2023 founders.

Business model

Self-serve software product with a free entry tier ("Start for free") plus paid and enterprise plans. Usage of AI-intensive actions is metered through a credit system: each account receives monthly credits that refresh automatically, with more intensive actions such as web search or data enrichment consuming more credits; enterprise plans offer higher credit limits, custom allocations, organization-wide playbook sharing and additional security features. The company also offers demos and direct contact for custom integrations and API access.

Subscription/plan-based access combined with credit consumption for AI-intensive actions, with an enterprise tier offering larger credit allocations and added security and sharing capabilities. One third-party profile estimates monthly revenue of about $417,000 and describes the company as profitable, but labels these figures as estimated from public sources rather than company-disclosed.

Traction

At the December 2023 launch the company was in private beta with a small number of customers and declined to share traction metrics. Its website later published named testimonials from users at organizations including Samsung, AMD, Mixrank, Brivity, Sonant, Darwin Ventures and Avelon Consulting, citing outcomes such as a process moving from 24-48 days to 12 hours and roughly 10x workflow throughput (customer-reported claims). A third-party profile lists an employee range of 51-200 and estimates monthly revenue of about $417,000 and profitability, though these figures are described as estimates derived from public sources.

Latest developments

Lutra's site now offers general self-serve sign-up rather than a waitlist, advertises SOC-2 compliance and enterprise features such as organization-wide Playbook sharing, and promotes MintMCP by Lutra for securely connecting enterprise data and applications to AI. The integration catalog has expanded across CRM, finance, productivity and data platforms, with custom connectors supported via MCP, REST and OpenAPI, and a further list of integrations announced as coming soon.