Nessie

What do they actually do

Nessie is a private‑beta Mac desktop app that imports your AI chat history (e.g., ChatGPT, Gemini, Claude), organizes it into topics and distilled notes, makes everything searchable, and lets you keep long conversations going past other platforms’ context limits—all in one personal knowledge base you can query and build on (nessielabs.com; YC profile).

The app is positioned as local‑first and privacy‑first: Nessie says it does not store the content of your chats or notes on its servers, and early posts describe local preprocessing with user control over what gets uploaded (nessielabs.com; Reddit post).

Who are their target customer(s)

• AI power users running long, evolving threads: They hit token limits and lose continuity, forcing them to manually stitch together context across multiple chat apps and documents.
• Researchers and writers iterating across many chats and drafts: Ideas and references are scattered across tools, making re‑finding prior insights slow and error‑prone; search is brittle and context gets lost.
• Founders and product managers tracking decisions over time: Reasoning and trade‑offs live in chat logs; there’s no single place to review how ideas evolved and reuse that context in current work.
• Privacy‑sensitive professionals (legal, medical, executives): They test or draft sensitive content with AI but don’t want raw chats stored on third‑party servers; they need a local‑first option with control.
• Multi‑platform AI users (ChatGPT, Gemini, Claude, etc.): Histories are fragmented and importing is manual or brittle; they want simple connectors plus unified search so prior work isn’t lost.

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

• First 10: High‑touch onboarding through YC and founders’ networks: personally help early users import their largest chat histories, fix connector issues live, and capture concrete before/after outcomes (nessielabs.com, YC).
• First 50: Expand to targeted communities (relevant subreddits, X/LinkedIn, YC alumni channels) with weekly live demos and office hours; codify common import flows into repeatable docs to reduce 1:1 time.
• First 100: Open a controlled public beta with self‑serve import guides and stabilized top connectors; run a few small team pilots (e.g., legal/research) to gather structured usage data and secure in‑product testimonials.

What is the rough total addressable market

Top-down context:

Nessie sits at the overlap of knowledge‑management software and AI assistants. Grand View Research estimates knowledge‑management software at about USD 20.15B in 2024 and growing, and the AI assistant market at ~USD 16.29B in 2024 and growing quickly (GVR KM report; GVR AI assistant report).

Bottom-up calculation:

If Nessie converts roughly 250,000 paying power users globally at ~$12/month (supported by widespread AI use among knowledge workers), that implies about $36M ARR as an initial SAM milestone (Microsoft Work Trend Index: 75% of knowledge workers use AI).

Assumptions:

• A sizable subset of AI‑active knowledge workers are willing to pay for a privacy‑first personal knowledge tool.
• 250k paying users is achievable with broader platform support and stable connectors over time.
• Average subscription pricing in the ~$10–$15/month range is sustainable for this category.

Who are some of their notable competitors

• Notion: All‑in‑one workspace used by individuals and teams; Notion AI adds Q&A and synthesis across docs. Broad, multi‑platform footprint and strong ecosystem.
• Mem: Personal knowledge app with AI features to capture, organize, and resurface notes and ideas; focuses on automated organization and retrieval.
• Obsidian: Local‑first markdown knowledge base with graph view and robust plugins (including AI). Appeals to privacy‑minded users who want full data control.
• Rewind: Mac app that records on‑device activity and makes it searchable; positioned as a personal memory layer and retrieval tool with a privacy posture.
• Tana: Networked note‑taking and knowledge graph app with AI‑assisted structuring and search; targets power users building complex personal knowledge systems.