Lotas

What do they actually do

Lotas builds AI‑native desktop tools for data science that run locally. Today they ship two downloadable apps: Rao, a packaged, modified RStudio with an embedded AI assistant, and Erdos, their own open‑source IDE (AGPLv3) built for data workflows (Rao, rao repo, Erdos, erdos repo).

In a typical session, the assistant reads your local project (files, notebooks, workspace), generates or edits code, runs it, and interprets console output, errors, and plots so you can iterate without leaving the IDE. The apps expose panels for console, plots, environment/variables, and rendered notebook/Rmd/Quarto output, so you and the AI see the same results (Rao changelog/features, Erdos features).

Users can choose secure, zero‑data‑retention backends or bring their own API keys; teams can deploy on‑prem for stricter compliance. Pricing is a Free tier (50 queries/month) and Pro at $20/month (500 queries/month, usage‑based overage), with enterprise options for teams (pricing).

Who are their target customer(s)

• Individual R users in RStudio (academics, researchers, analysts): They want code help that understands their local project and can run code and show plots without copying data into web services; Rao provides this as a downloadable RStudio fork with an embedded assistant (Rao, YC).
• Data scientists working across R, Python, and notebooks: They lose time switching tools and need one place that shows plots, variables, and notebooks together; Erdos aims to be a single, multi‑language IDE for this workflow (Erdos).
• Engineers/analysts with sensitive or regulated data: They cannot send data to third‑party APIs and need on‑prem or zero‑retention options and BYO keys; Lotas offers secure modes and enterprise/on‑prem deployment (pricing, Erdos).
• Early‑stage product/analytics teams: They spend cycles on boilerplate, debugging, and re‑runs; Lotas reads local files, runs code, and interprets errors/plots to shorten iteration (Rao, Erdos).
• IT/Procurement at larger organizations: They need predictable billing, deployment controls, and the ability to host internally or enforce API key policies; Lotas lists team/enterprise and on‑prem options (pricing).

How would they acquire their first 10, 50, and 100 customers

• First 10: Onboard known RStudio power users (academics, collaborators, active GitHub contributors) with hands‑on installs of Rao, gather direct feedback, and turn early users into public case studies and contributors (Rao, rao repo).
• First 50: Publish reproducible Rmd/Quarto examples and releases across GitHub and R/data‑science communities to drive organic installs; offer short Pro trials and support in exchange for structured feedback and testimonials (Erdos, erdos repo).
• First 100: Target small teams at startups and research groups with on‑prem pilots and security reviews; use published tiers and an enterprise pilot package to close paid seats and capture procurement contacts (pricing, Erdos enterprise notes).

What is the rough total addressable market

Top-down context:

YC cites an RStudio audience of ~5 million; at Lotas Pro pricing of $20/month ($240/year), that’s a clean single‑seat ceiling of ~$1.2B/year (YC, pricing).

Bottom-up calculation:

At $240/year, 0.5%–5% conversion of 5M RStudio users implies ~$6M–$60M ARR; reaching every 245,900 U.S. data scientist would be ~${59}M/year at the same ARPU (YC, pricing, BLS).

Assumptions:

• ARPU uses Lotas Pro at $240/year (pricing).
• Initial addressable base is ~5M RStudio users per YC (YC).
• Illustrative scenarios assume 0.5%–5% conversion of that base; enterprise deals are additional.

Who are some of their notable competitors

• Posit (RStudio): Incumbent IDE for R users with desktop, server, and cloud products plus paid team/enterprise offerings; it’s the default environment Lotas aims to replace or augment for R workflows (Posit/RStudio).
• Cursor: AI‑first coding IDE with agentic workflows and deep codebase indexing, including enterprise features and model choices; competes on the AI‑native developer experience (Cursor).
• JetBrains DataSpell: Desktop IDE for data analysis (Python, SQL, R) with integrated AI and database tooling; targets analysts who want a polished local environment (DataSpell).
• Deepnote: Cloud‑first collaborative notebook platform with AI features, scheduling, and enterprise security; competes on team collaboration and managed compute (Deepnote).
• Domino Data Lab: Enterprise data science platform emphasizing on‑prem/self‑managed deployments, governance, and reproducibility; strong fit for regulated customers needing audited workflows and enterprise controls (Domino).