Import your book, install the skills you need, and put your research on a schedule. Lyrra is one environment for portfolio monitoring, stock and sector research, and news alerts triggered by what's actually in your portfolio.
Connect your book. Write the job once. It runs forever.
Early cohorts get onboarding from the founders. No card required.
Chat assistants give you an answer and forget you exist. Dashboards show you numbers but can't reason about them. Terminals cost more than a junior analyst and still have no idea what's in your portfolio. None of them can do the work twice.
So the work stays manual. You already know how you evaluate a name — the checklist, the screens you re-run, the five things you always check before you size up. But that process lives in your head, one spreadsheet, and fourteen browser tabs, and every time you use it you rebuild it by hand.
Which is why your real coverage is however many names you can personally get to. Automate the process and that ceiling moves: the same rigor you apply to your top holdings runs across everything you follow, whether or not you had time this week.
Lyrra is the layer nobody built: somewhere to write your process down once and let it run.
Start with the spreadsheet you already keep. Drop in an Excel or CSV of your holdings — tickers, cost basis, weights, whatever columns you use — and Lyrra builds your portfolio graph from it. Connect a broker or custodian when you're ready. Everything downstream is scoped to what you actually own.
Skills are the units of work: portfolio monitoring, stock research, sector research, news coverage. Install the ones you need and point them at your mandate. Each one inherits your context — your positions, your watchlist, your notes, your constraints — so nothing has to be re-explained.
Every skill takes a schedule. Monitor the book at 7:00 on weekdays. Re-run the sector file the morning after CPI. Sweep news across the watchlist every four hours. Set it once and it happens whether or not you remember to open the app.
Skills are how work gets done in Lyrra. Each one is a repeatable job with your context baked in — install it, schedule it, and read the output.
Upload your Excel and Lyrra starts watching. Weight drift, concentration creep, P&L attribution, and anything in the book that moved for a reason you should probably know about. The output isn't a number that changed — it's what changed and why.
Point it at a ticker and get a structured writeup: the business, the numbers, the competitive position, the valuation, and the bear case. Built from filings, transcripts, estimates, ownership data, and anything you've uploaded yourself. Every claim carries its source so you can check the work instead of trusting it.
Constituents, share shifts, margin structure, capex cycles, regulatory overhang, and an honest read on who's actually winning. Same framework every time — which means sectors become comparable to each other, not just to themselves.
Lyrra watches your holdings and your watchlist, reads what breaks, and tells you the one thing that matters for your position. Not a digest of everything. A short answer to the only question worth asking: does this change the thesis?
Everything above is also a conversation. Ask Lyrra a question the way you'd ask a good analyst — about a position, a screen, a scenario, a number you don't believe — and it works from the same context the scheduled jobs use. It can pull data, run the analysis, build the chart, and show you its work.
Ask it to schedule something and it writes the job for you.
Lyrra connects through the Model Context Protocol, so your entitlements stay yours and your seat stays yours. We don't resell your data back to you — we read through the vendors you've already chosen.
EVERYTHING ELSE — MCP is an open protocol. If your vendor speaks it, Lyrra connects to it — and if they don't, tell us and we'll build it.
Developers stopped stitching together a text editor, a compiler, a debugger, and a terminal about thirty years ago. They got one environment where the pieces knew about each other, and their output went up permanently. Nobody argues about it now.
Investment research never got one. The data sits with a vendor. The model sits in Excel. The notes sit in Notion. The news sits in your inbox. The reasoning sits in your head, and none of it knows the rest exists.
Lyrra is the desk. Your book, your documents, your data vendors, your process, and an agent that can operate all of it — in one place, on a schedule you control.
Lyrra reads positions. It cannot place a trade, move cash, or change anything in your accounts. There is no code path that does.
Your portfolio, your documents, and your questions never train a model. Contractually, with every provider we use.
Your workspace is walled off from everyone else's. Jobs execute in a sandbox scoped to you and nobody else.
Disconnect and your data goes with you. Export everything, delete everything, no phone call required.
Encryption at rest and in transit. SOC 2 Type II work underway — happy to share where we are.
No. Lyrra is research infrastructure. It produces analysis; you make every decision.
No. All portfolio connections are read-only, by design.
For premium sources like FactSet or Capital IQ, yes — Lyrra reads through your entitlements. Public sources like EDGAR and transcripts are included.
That's the intended starting point. Excel import works on day one and everything else is optional.
No. Describe the schedule in plain English and Lyrra writes it.
Cohort 01 opens this quarter. Waitlist position is weighted by fit, so the free-text field is worth filling in.
We're taking a small number of investors at a time, because the skills only get good if we build them next to the people using them. Tell us what you run and what you'd automate first. The sharpest answers get in soonest.
We read every submission. No drip sequence, no newsletter — one email when your cohort opens.