Notes that index your thinking.
I spent years as an analyst at a hedge fund, and most of what annoyed me about the job wasn't the thinking. It was everything around the thinking. Basis started as my attempt to delete that layer of the job.
Six tabs and a prayer
My actual workflow looked like this: filings and transcripts in BamSEC, prices in the Stocks app, fundamentals and news on a Bloomberg, notes in Evernote, documents loaded into NotebookLM when I wanted AI to read something, and ChatGPT open in another tab for everything else. Six good tools. Zero shared memory.
The AI tools in particular were incredible, and incredibly frustrating. The model could reason brilliantly about whatever I pasted into it, but it had no idea what I was working on. It hadn't read my notes. It didn't know my coverage, my thesis, or what management promised last quarter. Every conversation started from zero, and the burden of assembling context (copying passages, re-explaining positions, re-uploading documents) fell on me. Manually. Every time.
The fix wasn't a better model. It was a better substrate.
What I actually needed was for all of it to live in one place, so the AI could synthesize and act on it instead of me playing courier between tabs. So I built that, one piece at a time:
- A listening engine that watches news, filings, transcripts, social chatter, and price action for the names I track, and filters it against what I actually care about.
- A note-taking service I'd want anyway: a first-class editor with ticker pills, contact tags, and a focus mode, so a note written today is findable and linkable years from now.
- Prices and fundamentals: live quotes, history, estimates, and financials on every name, in the same workspace as the research about them.
- An AI agent on top of all of it: one that reasons over my notes, the transcripts, the feed, and the market data together, because they finally share one home.
That last point is the whole product. Notes stop being a filing cabinet and become an index of your thinking: the thing the agent reads before it answers, the memory every new piece of information gets weighed against.
What centralized context unlocks
Once the data and the thinking live together, the investing-specific use cases get sharp: ask what management said about capacity and whether your own notes believed them. Get a morning brief ranked against your holdings, not a generic feed. Have an earnings call graded against the expectations you wrote down before the print. None of that is possible when the context is scattered across six apps that have never heard of each other.
I built Basis because I wanted it to exist. If your desk looks like mine did (brilliant tools, none of them talking), it might be what you wanted too.