Products

Software the lab ships.

Small, opinionated tools that started as our own itches. Some are usable today; some are still finding their shape. We list the honest status next to each one.

HiringSignal Shipping soon

Real hiring managers, not recruiters.

LinkedIn is full of hiring posts. Most are recruiters, agencies, and reshares. HiringSignal ranks posts by who’s actually hiring — founders, engineering managers, CTOs — and drops the rest. Each post scores 0–100 on linguistic signals: hiring-manager titles, first-person voice, direct CTAs, stack specifics. Recruiter, agency, and reshare patterns subtract.

  • Vertex AI Search
  • Rule-based scorer
  • CLI & landing page
  • Tunable term lists
On GitHub
Altassian Open source

An open-source Confluence alternative.

A wiki platform for teams to create, organize, and collaborate on documentation — without the enterprise tax. Markdown-based pages, hierarchical structure, spaces, versioning, comments, attachments, and full-text search. Self-hostable. Built for teams who want a documentation tool they actually own.

  • Django + DRF backend
  • SvelteKit frontend
  • JWT auth
  • Page versioning & history
  • MIT-licensed
On GitHub
Project Pager In development

A phone-first surface for AI coding agents.

When you walk away from your laptop, your AI coding agents shouldn’t be stuck behind your terminal. Project Pager lets you continue chatting with every active agent — Claude Code, Codex CLI, Cursor, anything that speaks MCP — from a single phone app. Discrete sessions, rich rendering of diffs and tool calls, first-class file uploads. Designed for power users who run multiple agents at once.

  • MCP server on the agent side
  • Cloud relay
  • Phone app, one inbox
  • Approval-prompt rendering
On GitHub
AgentPass Concept

1Password for AI agents.

A credential broker that lets AI agents log in to the services they need, without ever seeing your passwords. The agent receives short-lived capabilities scoped to the work at hand; the broker handles auth, rotation, audit, and revocation. Designed around the principle that agents should hold capabilities, not secrets.

  • Capability-scoped tokens
  • Short-lived sessions
  • Audit log per action
  • Revocation surface
On GitHub

Working on something adjacent?
We should talk.

Start a conversation