ABAI Berkshire Space

AI investment research planner

Make every stock thesis survive four masters before capital is at risk.

Turn a stock idea into a disciplined research plan with four-master tension, exact-data checkpoints, quick-kill red lines, and a paid memo workflow that refuses vague conviction.

View pricing packages Preview planner

Independent companion for the open-source AI Berkshire framework. Educational workflow support only, not financial advice.

5,984 GitHub stars 791 forks MIT source license 18 research skills 74 example reports

A research workflow, not a stock tip machine

AI Berkshire Space helps investors plan research before they ask an AI model for a verdict. The workflow forces a company through four perspectives, data-validation checks, inversion, and post-buy falsifier tracking.

The public preview shows the shape of the memo. Paid access unlocks structured brief generation and checklist output, while runtime APIs record paid-gate, checkout, and planner events when D1 is configured.

Sample preview

Suggested path: circle of competence, business quality, moat durability, management incentives, valuation sanity, inversion, quick-kill red lines, and a falsifier list. Full briefs unlock after package selection.

What the planner builds

Four-master tension

Buffett-style valuation, Munger-style inversion, Duan Yongping business quality, and Li Lu long-term certainty are kept separate before synthesis.

Financial rigor

Market cap, valuation, cross-source data, and report-audit checks are treated as gates, not decoration.

Decision discipline

Quick-kill rules, gray-zone labels, mirror-test summaries, and thesis falsifiers keep the memo auditable.

Source boundary

Independent

This site is not official and is not endorsed by the upstream maintainer.

Source-derived

Facts come from the public repository, local source scan, MIT license, skills, prompts, tools, and report examples.

Risk-aware

No public page promises returns, future performance, or buy/sell outcomes. Historical examples remain source facts, not site traction.

Packages before output

Functional research brief generation is gated. Users can read the method and preview a sample, then choose a package before a full memo is generated.

Open pricing

Starter

One research memo outline for a single public-company idea.

$4.50/mo

$54 due today

  • 1 paid research brief
  • Four-master lens map
  • Quick-kill checklist
  • Sample evidence plan

Pro

Best fit

A reusable research pack for investors comparing multiple theses.

$14.50/mo

$174 due today

  • 5 paid research briefs
  • Financial rigor checklist
  • Falsifier tracker
  • Portfolio review prompts

Enterprise

Team-ready planning for repeatable research review and audit trails.

$29.50/mo

$354 due today

  • 15 paid research briefs
  • Evidence matrix
  • Reviewer workflow
  • Private support intake

Payments are one-time and do not automatically renew. Annual pricing covers one year at a 50% lower monthly equivalent.