Open source · AI Labs · The Lossless Group
Multi-agent investment memo orchestrator.
Turn a company name into an institutional-quality investment memo with 33 specialized agents, premium data sources, full citation trails, and firm-scoped branded exports.
Analyst + MemoPop = 1–2 hours to a polished, cited, firm-branded investment memo.
Why this exists
Asking ChatGPT to "write an investment memo" produces generic, uncited, surface-level output. Real memos need structured research from authoritative sources, firm-specific evaluation criteria, and analytical rigor that no single prompt can deliver.
A quality memo takes 2–5 days of research, writing, and revision — time most emerging managers don't have. Most of that work is pulling Crunchbase data, sizing TAM from Statista, cross-referencing team backgrounds, sourcing experts… tasks AI agents can do in minutes.
Every fund evaluates deals differently. Generic AI can't reflect your IC's actual decision framework. Scorecards and firm-scoped templates make your proprietary criteria the backbone of every memo.
The pipeline
Specialized agents, each producing research artifacts that serve as a source of truth, full retraceability, and a fallback for future iterations.
Thorough dataroom object identification and deconstruction. Pull KPI timelines, key charts, properly identify slide messages, and extract primary source information. Classify 12+ document types with confidence scoring.
Multi-query research via Perplexity Sonar Pro with section-specific @source targeting. Crunchbase, PitchBook, Statista, CB Insights.
Proprietary document structure loaded before writing begins. Guiding questions, vocabulary and phrasing, dos and don'ts — your firm's voice and analytical framework baked into every section.
Individual sections drafted by specialized writing agents shaped by the loaded outline. Each section processed independently to avoid timeouts.
Trademark links, social profiles, citations, competitive context. Each enrichment agent works on section files, not assembled content.
Three-step extract → verify (Perplexity) → correct (Claude) pipeline. Full traceability for every claim.
Firm-specific evaluation criteria applied after writing. Team velocity, market timing, defensibility — whatever your IC actually debates. Scores and flags feed back into revision.
Quality scoring against analytical standards. Threshold of 8/10 for auto-finalization. Detailed improvement suggestions below threshold.
Branded HTML/PDF in light and dark modes. Firm-scoped configs for colors, fonts, logos. One-pager summaries.
One command
$ memopop
╭─────────────────────────────────────────────────────────╮
│ │
│ Investment Memo Orchestrator │
│ v0.5.3 · 33 agents · Powered by Claude + Perplexity │
│ │
╰─────────────────────────────────────────────────────────╯
? What would you like to do?
❯ Generate a new investment memo
Generate a one-pager summary
Export an existing memo (HTML / PDF / Word)
Improve a specific section
Integrate content from versions
Run a specific agentNo flags, no file paths. The interactive CLI auto-discovers firms, deals, and versions.
MemoPop runs on your machine with your API keys. Claude for analysis and writing, Perplexity Sonar Pro for cited research, or DuckDuckGo as a free fallback. Your deal data stays yours.
Hit us up. We can turn around a draft same day, sometimes in 10–15 minutes.