AAIF Ambassador

ThreadWeaver
Continuity Between Thinking Spaces and Build Spaces

ThreadWeaver is designed to bridge ChatGPT ideation history with active Goose workflows. It exists to carry context across tools so practitioners can keep strategic continuity without manual copy-paste loops.

GitHub @threadweaver/cli @threadweaver/sdk Live Docs

The Problem

Why ThreadWeaver Was Built

A large share of ideation and strategic thinking happens in ChatGPT before implementation begins in Goose. There is no native bridge for referencing that prior reasoning inside active Goose sessions.

In practice, that created repetitive workflow friction: copying and pasting long chats, or repeatedly asking ChatGPT to export discussions into markdown files for manual handoff into current Goose work.

ThreadWeaver addresses this by turning chat history into a retrievable, project-scoped memory layer so prior thinking can inform current execution without breaking flow.

Design Choices

1) Continuity as the Core Primitive

The core design target is continuity between separate thinking spaces/tools and active build sessions. ThreadWeaver treats historical ideation as reusable working context rather than static archive material.

2) Goose-First, Conversation-Native UX

ThreadWeaver is built for Goose and works with Goose plus ChatGPT data exports today. Queries are intentionally natural-language so retrieval can happen inside normal Goose collaboration loops.

3) Consent and Scope Controls

Project-scoped access and explicit allow rules are prioritized so retrieval stays bounded. The goal is useful continuity without losing user control over memory boundaries.

4) Local-First with Portable Direction

Current releases optimize for local, pragmatic usage with ChatGPT exports. Upcoming releases are maturing toward broader portability so additional tools and contexts can participate through cleaner interfaces.

How It Works Today

Point Goose + ThreadWeaver at Your ChatGPT Export

  1. Request a ChatGPT export and wait for the download email.
  2. Download and unzip the export locally.
  3. In Goose, provide the absolute path to the unzipped export directory.
  4. Run the ThreadWeaver commands below to connect and index your export data.

Export instructions: How to export ChatGPT history and data 

threadweaver connect chatgpt-export --file /absolute/path/to/chatgpt-export
threadweaver projects sync-from-import
threadweaver projects allow --project chatgpt-general --access summary

Use It Inside Goose

After setup, ask naturally in Goose to pull prior ideation into the active build flow:

I remember thinking in ChatGPT about X.
Surface that prior thinking and connect it to what we are building on project Y.

AAIF Ecosystem Alignment

Aligned with Goose's Open Direction

Angie Jones' update, Goose doubles down on open in latest two releases, highlights a direction toward open interfaces, ecosystem interoperability, and composable agent workflows.

ThreadWeaver aligns with that trajectory by focusing on reusable memory interfaces, policy-governed retrieval boundaries, and practical interoperability between where practitioners think (ChatGPT) and where they execute (Goose).

Today the integration path is Goose + ChatGPT export. The roadmap is focused on expanding portability so continuity can move across more tools and practitioner workflows in upcoming releases.