Brad Zhang / public archive

Brad Zhang

AI product notes, agent workflow writing, and open-source dossiers for founders and early technical teams.

Directory / longform / forum / open source

Brad Zhang: archive, notes, and projects.

This is a working table of contents for public writing and shipped work. Move directly into the archive, the forum, the proof library, or the collaboration routes.

Current issue

A reading order through the public proof system.

Open thesis map

Directory 01 / Longform

Longform archive

Essays, articles, and long X posts. This is the best place to understand the underlying arguments.

Selected entries / 0 of 0

Browse all

Directory 02 / Forum

Filtered X forum

Shorter posts that still carry information. Low-signal replies, pure acknowledgements, and lightweight chatter are filtered out.

Dispatch index / 0 of 0

Open board

Directory 03 / Open source

Open-source hall

Public repositories, README rendering, and founder-readable project dossiers.

Featured repositories / 5 of 22

View all

01 / Python

7.9k stars

Lead dossier

fireworks-tech-graph

A project that treats architecture diagrams not as cosmetic add-ons, but as durable explanatory surfaces for AI...

This repo is market proof that the builder can turn vague technical structure into public art...

Lead dossier entry

Dossier page / fireworks-tech-graph

Proof continuation

Founder-readable flagship case study

Case page / flagship proof

Founder fit

Founder collaboration fit

Fit route / founder collaboration

Dossier route

02 / Python

94 stars

Lead dossier

fireworks-skill-memory

A memory system for coding agents that values persistence, reuse, and editorial recall over mere accumulation.

This is one of the clearest proofs on the site that AI memory should be structured, selective...

Lead dossier entry

Dossier page / fireworks-skill-memory

Proof continuation

Wider thesis and proof graph

Thesis route / proof graph

Founder fit

Founder collaboration fit

Fit route / founder collaboration

Dossier route

03 / Python

8 stars

Lead dossier

fireworks-sessions-saver

A checkpointing layer for coding sessions, built to survive interruption, context loss, and handoff without pret...

Interruption is not an edge case in real AI work. A team that can resume cleanly has a direct...

Lead dossier entry

Dossier page / fireworks-sessions-saver

Proof continuation

Harness engineering for agentic AI products thesis-side continuation

Cluster route / thesis map

Founder fit

Founder collaboration fit

Fit route / founder collaboration

Dossier route

04 / TypeScript

2 stars

Lead dossier

paper-pal

A paper-reading tool that frames comprehension as an active workflow of margin notes, synthesis, and return visits.

Serious AI work depends on reading systems, not just browsing papers. This project treats com...

Lead dossier entry

Dossier page / paper-pal

Proof continuation

Wider thesis and proof graph

Thesis route / proof graph

Founder fit

Founder collaboration fit

Fit route / founder collaboration

Dossier route

05 / Python

390 stars

Lead dossier

media-downloader

🎬 智能媒体下载器 - 根据用户描述自动搜索和下载图片、视频片段 | Smart Media Downloader - Auto search and download images/videos based on des...

Creative and technical distribution systems both break when asset collection stays manual and...

Lead dossier entry

Dossier page / media-downloader

Proof continuation

Technical distribution for AI builders thesis-side continuation

Cluster route / thesis map

Founder fit

Founder collaboration fit

Fit route / founder collaboration

Dossier route

Python / 7.9k stars

fireworks-tech-graph

A project that treats architecture diagrams not as cosmetic add-ons, but as durable explanatory surfaces for AI systems work.

This repo is market proof that the builder can turn vague technical structure into public artifacts developers want to share.

Python / 94 stars

fireworks-skill-memory

A memory system for coding agents that values persistence, reuse, and editorial recall over mere accumulation.

This is one of the clearest proofs on the site that AI memory should be structured, selective, and recoverable rather than merely larger.

Python / 8 stars

fireworks-sessions-saver

A checkpointing layer for coding sessions, built to survive interruption, context loss, and handoff without pretending memory never breaks.

Interruption is not an edge case in real AI work. A team that can resume cleanly has a direct execution advantage.

TypeScript / 2 stars

paper-pal

A paper-reading tool that frames comprehension as an active workflow of margin notes, synthesis, and return visits.

Serious AI work depends on reading systems, not just browsing papers. This project treats comprehension as production work.

Python / 390 stars

media-downloader

🎬 智能媒体下载器 - 根据用户描述自动搜索和下载图片、视频片段 | Smart Media Downloader - Auto search and download images/videos based on description

Creative and technical distribution systems both break when asset collection stays manual and scattered.

Python / 165 stars

skills-updater

Manage, update, and discover Claude Code skills. 管理、更新和发现 Claude Code 技能。

Reusable AI workflows decay unless updates and discovery stay easy. Maintenance tooling is what keeps reuse real.

Next step

Collaboration, fit, and ways of working.

Work with me

High-trust lane 01

Founding engineer fit

Early AI teams that need product-minded engineering around agent workflows or developer tools.

Open-source traction, shipped product surfaces, X field notes, and a founder-readable site.

A concrete trial module with a repo, demo, docs, or workflow to inspect.

High-trust lane 02

Design partner sprint

Teams with one fragile AI workflow that should become reliable, inspectable, and reusable.

The AI Devtool Sprint offer, proof clusters, and case-study style artifacts.

A two-week paid sprint around one artifact and one handoff.

High-trust lane 03

Developer-facing proof work

AI infra, devtool, and internal-copilot teams whose product is strong but hard to explain.

fireworks-tech-graph, project dossiers, technical diagrams, and launch narrative systems.

Package the capability into docs, diagrams, examples, and a public proof surface.