Second-Inc

Second-Inc / second

Public

The factory for custom internal software, purpose-built for human2agent work.

docs.second.so agent-management agents ai-agents internal-tool nextjs
13
0
89% credibility
Found May 19, 2026 at 18 stars 3x -- GitGems finds repos before they trend. Get early access to the next one.
Sign Up Free
AI Analysis
TypeScript
AI Summary

Second is an open-source platform that lets teams build custom internal tools where humans and AI agents work together on the same interface. Instead of adapting workflows to pre-built software, you describe what you need and Second generates a full internal app with a real-time database, permissions, and audit logs — all running on your own infrastructure. It supports multiple AI runtimes (Claude Code, Codex, OpenCode), includes enterprise-grade security with zero-trust agent controls, and can be deployed on Kubernetes or run locally. Think of it as an internal, secure, collaborative alternative to no-code app builders — purpose-built for long-running, asynchronous work with AI agents.

How It Works

1
🏢 Set up your workspace

You create your company workspace with a name that anchors everything you build together.

2
👤 Introduce yourself

You enter your name and role so the system can personalize your experience and give you relevant context.

3
🤖 Connect your AI assistant

You connect your preferred AI assistant (Claude Code, Codex, or OpenCode) so it can help build your apps.

4
✨ Describe what you need

You tell Second what kind of internal tool you want — like a competitor research tracker or lead enrichment pipeline — and your AI assistant builds it for you.

5
Two paths emerge
👤
Use it yourself

You start using the app right away, with your AI assistant working alongside you on the same interface.

👥
Invite your team

You add teammates to your workspace, assign roles, and set up approval workflows before publishing.

6
🔒 Everything stays secure

Your app runs behind your own firewall with full audit logs, role-based permissions, and zero-trust agent controls.

🎉 Your team builds together

Humans and AI agents collaborate on the same custom interface, sharing real-time data without copy-pasting or context switching.

Sign up to see the full architecture

5 more

Sign Up Free

Star Growth

See how this repo grew from 18 to 13 stars Sign Up Free
Repurpose This Repo

Repurpose is a Pro feature

Generate ready-to-use prompts for X threads, LinkedIn posts, blog posts, YouTube scripts, and more -- with full repo context baked in.

Unlock Repurpose
AI-Generated Review

What is Second?

Second is an open-source platform for building internal software where humans and AI agents work together in the same workspace. Built in TypeScript with Next.js, it lets teams create custom apps through prompts while agents handle long-running async tasks like research, enrichment, and monitoring. The key difference from typical AI tooling: agents read and write to the same real-time database your team uses, with scoped permissions and full audit trails. You bring your own agent runtime (Claude Code, Codex, or OpenCode), and Second handles the governance, integrations, and collaboration layer around it.

Why is it gaining traction?

The hook is treating agents as first-class citizens inside the apps you build, not as side conversations. Companies like Ramp and Deel built internal platforms to solve this exact problem. Second gives every organization that capability. The governance model is particularly compelling: agents.json declares exactly what tools and data each agent can access, with approval workflows before anything goes live. Secrets stay server-side, private IPs are blocked, and every action gets logged. The self-hosting story is strong for security-conscious teams.

Who should use this?

Platform engineers at companies with Claude or Codex prototypes that never reached production due to security, governance, or integration headaches. Teams building internal tools where humans and agents collaborate on workflows like competitor research, lead enrichment, or compliance monitoring. Organizations with strict data residency or security requirements that need on-prem deployment with full control.

Verdict

Second solves a real problem that increasingly matters as AI agents move from prototypes to production. The credibility score of 0.8999999761581421% reflects an early-stage project with only 13 stars, so treat it as promising but immature. The documentation is solid and the architecture is well-thought-out, but test coverage and community size are unknowns. Worth evaluating if you need human-agent collaborative internal tools; wait for more community adoption if you prefer battle-tested solutions.

Sign up to read the full AI review Sign Up Free

Similar repos coming soon.