Skip to content

Latest commit

 

History

1 Commit

Folders and files

NameName
Last commit message
Last commit date
 
 
 
 
 
 
 
 

Repository files navigation

OpenClaw · Equity Research Agent

An 8+1 AI research team for sell-side and buy-side analysts ── delivering auditable research drafts, not chat.

This repository is the open product spec for OpenClaw's first vertical flagship — the Equity Research Agent. We're publishing it openly so prospective design partners can co-shape the product before V1 freezes.

This is not a code repository (yet). It's a living PRD, a problem statement, and an invitation.


Why this exists

Every major equity research workflow today is either:

  • Manual & slow (analysts spreading filings into Excel at 11pm on earnings day), or
  • Tool-assisted but ungrounded (LLM summaries with no citations, no audit trail, no compliance gate).

Existing players (AlphaSense, Hebbia, Daloopa, Bloomberg GPT, FinChat) deliver search + summarization. None deliver team-based, citation-grounded, audit-ready research drafts.

OpenClaw Equity Research Agent is that missing layer.


What's in this repo

File Purpose
PRD.md The full v0.1 product spec (19 sections, ~27 KB)
CONTRIBUTING.md How design partners and reviewers can shape the spec
docs/ Deep-dive supplements (coming as v0.2 evolves)
.github/ISSUE_TEMPLATE/ Structured slots for feedback, questions, gaps

The 8+1 Agent Team (at a glance)

                  +1 PM Orchestrator (FDE)
                  覆盖优先级 · 发布日历 · 质量把关
   ┌─────────────────────────────────────────────────┐
   │  INTELLIGENCE  ─  Filings Hunter                 │
   │                ─  Transcript Listener            │
   │                ─  Macro / Sector Scout           │
   │  ANALYSIS      ─  Modeler                        │
   │                ─  Comp Builder                   │
   │  SYNTHESIS     ─  Thesis Drafter                 │
   │                ─  Report Writer                  │
   │  GATE          ─  Karen · Compliance / Truth Gate│
   └─────────────────────────────────────────────────┘

Each agent has its own responsibility, KPI, escalation rule, and — critically — its own citation contract.

See PRD.md §8 for the full per-agent spec.


Core design principles

  1. The AI never proposes rating / target price. That's the analyst's accountable decision. AI surfaces sensitivities; humans make calls.
  2. Every claim is citation-bound. No citation, no publish. Karen / Truth Gate enforces.
  3. Math runs in deterministic code, not LLMs. DCF, multiples, growth — Python functions, never generated tokens.
  4. MNPI walls are first-class. Data is tagged on ingest; agents inherit persona permissions; tries to cross logged immutably.
  5. Earnings cycle first. Pre-earnings preview + post-earnings recap is the highest-frequency, most-repetitive, most-quantifiable workflow. Start there, not at initiation reports.

What we're looking for from design partners

You are... What we'd love to learn
Single-PM hedge fund / family office What does your earnings-night workflow actually look like? What would 4 hours saved mean?
Buy-side research head (5-20 analysts) Where do junior analyst hours leak? What template formats do you live in?
Sell-side analyst / compliance officer What does your compliance approval gate require? What would Karen need to clear it?
Corporate IR / strategy / PE post-investment What of this generalizes to your adjacent workflow?

How to engage

Comment on the PRD

Open an issue using one of the templates:

Become a design partner

Email hello@agentify.ai with:

  • Your firm type & size
  • Your coverage scope (sectors / geographies / companies)
  • 30-min slot you can do this week

Design partners get 6 months free use in exchange for:

  • 1 hour / week of feedback during V0-V1
  • Permission to be quoted in launch case study (under NDA terms you set)
  • Honest pushback when we're wrong

What this is NOT

  • Not a Bloomberg replacement. We complement market data terminals.
  • Not an alpha-generation engine. No "what stocks to buy" magic.
  • Not auto-publish. Always human-in-the-loop. No exceptions.
  • Not open-source code. The PRD is open; the implementation will be commercial. (We may open-source components later — e.g., the citation framework — but this isn't the place.)

Roadmap

Phase Window Milestones
V0 · Stealth W1–W8 1 design partner · A1+A4+A7 manual end-to-end
V1 · Design Partner W9–W24 Full 8+1 · 3 partners · 2 earnings seasons
V2 · GA W25–W52 FactSet integration · sell-side POC × 1 · ARR ¥3M
V3 · Scale Y2 Alt-data · HK/SG expansion · ARR ¥15M

Full detail in PRD.md §14.


About OpenClaw / Agentify AI

OpenClaw is the multi-agent orchestration engine that powers Agentify AI. Agentify AI delivers AI job teams to enterprises — not tools, not seats, but teams that show up to work.

The Equity Research Agent is the first vertical flagship built on OpenClaw, following the 8+1 pattern already proven in:

  • AI Product/R&D Team (engineering delivery)
  • AI Overseas Marketing Team (verified marketing execution)

If this pattern resonates for your workflow vertical, let's talk.


License

PRD content: CC BY-SA 4.0 — share, adapt, build on it, attribute. Code (when published): TBD (likely Business Source License or proprietary, depending on component).


Contact

Build with us, not for us. The best PRDs are co-authored.

About

Open product spec for OpenClaw's Equity Research Agent — an 8+1 AI research team for sell-side/buy-side analysts. Co-design welcomed from design partners.

Topics

Resources

Contributing

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors