Agent-First Kit

Your AI can think.
Now let it do.

Agent-First Kit gives your AI agent a small set of tools — to read your email, open the real web, query a database, move a payment — that run right on your own computer and hand back clean, reliable results instead of guesses. You stay in control: nothing is sent, changed, or paid until you say so.

Human, agent, tools, and world arranged left to right: a small human node connects through a thick channel to a hexagonal agent core, which fans out to a cluster of distinct tool shapes, which in turn radiate threads to a field of varied world nodes.
The kit

What your agent can do

Each tool runs locally and works from real results, not guesses. Free and open source.

Agent-First Data

convention & protocol

A naming convention that lets AI agents understand your data without being told what it means, plus a CLI and library for reading and safely editing structured JSON, TOML, YAML, dotenv, and INI documents.

Ask your agent

“Apply the Agent-First Data convention across my project's fields, config, and logs.”

Agent-First HTTP

CLI tool

Give your AI agent its own private browser — so it reads the real page, past logins and bot walls, without ever touching yours.

Ask your agent

“Open this link and pull the pricing table out for me.”

Agent-First PSQL

CLI tool

A PostgreSQL interface for AI agents: reliable, structured, explicit, and read-only by default.

Ask your agent

“How many orders shipped late last month?”

Agent-First Pay

CLI tool

A payment tool for AI agents — send and receive across five networks through one interface, with spending limits you control.

Ask your agent

“Pay this $20 invoice from my Lightning balance.”

Agent-First Mail

CLI tool

Let your AI agent work your inbox — email pulled into plain files it reads, sorts, and drafts on your machine, with nothing sent until you confirm.

Ask your agent

“Sort this week's inbox and draft replies to anything urgent.”

Agent-First UI

UI protocol

A protocol that lets a tool publish what is true and what can be done as plain data, into a shared live space — so any viewer you trust, human or agent, renders the interface, and nothing the tool publishes can run on your machine. One tool, any viewer: the same space drives a web workbench, a terminal interface, a native app, a voice assistant, or an agent operating it directly.

Agent-First Slug

Rust library

Rust slug generation with explicit caller configuration for path and URL path segments.

Ask your agent

“Add agent-first-slug to my project and use it to slugify titles for URL path segments.”

The idea

An agent changes how we use software

It doesn’t just answer questions — it does the work, weaving together however many tools and sources a task needs, with none of the learning curve you used to pay. You say what you want; the agent handles the how.

  • Software you no longer have to learn

    For decades, using a computer meant learning each app — its menus, its quirks, where every setting hides. With an agent you just say what you want in plain language, and it operates the software for you. The learning curve doesn’t shift onto you; it disappears.

  • But the software wasn’t built for agents

    When it works it feels effortless — but underneath, the agent is driving tools made for a human at a screen. So it improvises: scraping text, guessing what a value means, mistaking an error for content. That gap is where things quietly break.

  • That’s what Agent-First Kit is for

    We build the tools an agent actually needs to act in the real world — each one returns clean, structured results it can rely on instead of guessing. Some are everyday essentials rebuilt the right way; some are entirely new. A small kit today, and growing.

  • The power is in combining them

    One tool is useful; the real leverage is an agent chaining several — read the email, look up the order, issue the refund, write back. Agent-First Kit is built to compose: each tool does one thing well and works alongside whatever else your agent already has, ours or not.