' Intrane.fr — Human expertise. Machine scale.
Gateway for SMEs

Human expertise.
Machine scale.

Staff augmentation and AI automation — in one subscription.

For solo CTOs, technical founders, and teams who need more bandwidth.

90% of your infrastructure problems are solved by automation.
The other 10% need architecture depth. That's why every plan
pairs AutoMaintainer with direct access to a Senior Architect.

Staff Augmentation AI Automation Open Toolkit
Try AutoMaintainer free View Bundles Explore the Toolkit
90%
OSS
Free to Use
Agent
First
JSON Native
15+
Years
Shipping Code
Javier Leandro Arancibia

Javier Leandro Arancibia

Senior Architect · Technical Founder

📍 Savoie, France 🔧 github.com/javimosch ✉️ arancibiajav@gmail.com

Born in Argentina. Studied Software Engineering at UTN Mar del Plata and UBA Buenos Aires. Built a career across continents — Buenos Aires, Lima, Barcelona, Paris — from ERP systems to fullstack web development. Had a spot at Station F in 2017-2018, the world's largest startup campus.

AI since 2013. Back when that meant coding neural networks for games, not calling APIs. That early conviction shaped everything — the obsession with machine-first design, the respect for human craft, and the belief that the two work best together, not replacing each other.

Today I build agent-first infrastructure for SMEs at Intrane. Products first (free OSS tools), then myself when you need architecture depth. Based in Savoie, France. Working globally.

Flagship Service

AutoMaintainer

Loop engineering, productized. A fleet of AI agents runs a verified build→review→fix loop on your repos — across frontier & mid-tier models, on your keys.

Verified Gate

Every change runs through build → test → fix → re-verify. Only work that passed becomes a PR. No slop.

BYOK

Bring your own keys. You pay tokens directly to your AI provider with zero markup. Full control.

Scheduled Runs

Set a cron-style window or trigger instantly. Agents work while you sleep. PRs appear automatically.

Try AutoMaintainer free

Free forever · No credit card · You stay in control

Try AutoMaintainer free
Affordable Alternative

mago

Cheap autonomous agents that run companies. A single self-run worker drives a whole team of agents over your GitHub repo — €20/month flat, BYOK, CLI-only. The lean, DIY counterpart to AutoMaintainer.

€20/month flat

One plan, no tiers, no per-resource limits. Agents back off when idle, so an idle company costs almost nothing.

A full team, you're the CEO

Seeds an executive team — CTO, CMO, Head of Product, Head of Org Engineering — over your repo. You set direction by filing issues; they ship PRs.

Git-native & yours

Your GitHub repo is the company — issues are the backlog, PRs are deliverables, comments are the conversation. BYOK keys stay on your worker. git clone is the audit log.

Try mago

€20/month · BYOK, your keys never leave your worker · You own everything

The Intrane Toolkit

20+ open-source tools across every layer of your stack — from AI agent frameworks to DevOps utilities.

Flagship AI & CLI

SuperCLI

6,000+ tools. One CLI. Zero configuration. Agent-first CLI that unifies tools, APIs, and workflows into a single execution layer.

New Agent Ops

roam

Leave a working agent on a remote box. Dispatch a self-replicating agent to a VM, walk away, and approve its risky moves from your phone — one static binary, no Python on the target.

CLI Agents Tooling
6,000+
Tools
20+
Repos
View Source Try it now
€20/mo Autonomous Agents

mago

Cheap autonomous agents that run companies — the lean, self-run alternative to AutoMaintainer. One worker drives a CTO, CMO & product team over your GitHub repo. BYOK, CLI-only, your repo is the company.

Agents Git-native BYOK Go
Try mago Learn more →
€29/mo GTM & Leads

grepapi

An affordable, agent-first SerpApi alternative + GTM warm-lead engine. Point it at an ICP; get real companies with phone numbers and warm LinkedIn leads, each with a brief your own LLM turns into outreach. BYO-browser (no proxy bill), BYO-LLM (no token markup), CLI-only.

GTM Agent-first BYO-browser machin
Try grepapi Read the story →
NodeJS

Superbackend

The All-In-One Boilerplate. 20+ integrated systems including IAM, Headless CMS, and Audit Logs. Stop building infrastructure, start shipping features.

IAM CMS LLM Docker
View Source Docs →
AI & Automation

Superbot

A lightweight, open-source alternative to Clawdbot. Designed for fast AI agent orchestration and workflow automation without the bloat.

Agents Orchestration Webhooks
View Source Docs →
Analytics

Superinsights

Comprehensive analytics platform. Track user behavior, system performance, and business metrics in one unified dashboard.

View Source Docs →
DevOps & CMS

Superlandings

Manage and publish multiple landing pages with admin panel, Traefik custom domains, CLI, and Docker.

View Source Docs →
Messaging

Superqueues

A lightweight RabbitMQ wrapper gateway with HTTP API. Simplify message queue management for distributed systems.

View Source Docs →
Infrastructure

Supersettings

Multi-tenant settings microservice with dynamic auth. Centralized configuration management for complex SaaS architectures.

View Source Docs →
Live Agent-First

hart

The agent-first artifact host. Any terminal agent publishes self-contained HTML or JSX and gets a live, versioned, sandboxed URL — one self-hosted binary. Claude Artifacts, unbundled and owned.

View Source Live demo →

Agent Frameworks

Libraries and protocols purpose-built for agent-to-agent and agent-to-tool communication.

ContextFS

MCP

Persistent, structured filesystem for LLM agents. MCP-compatible, multi-tenant, zero-dependency dashboard. Enables agents to read, write, and search files safely without direct machine access.

View Source

a2a

AM Core
P2P Messaging

Peer-to-peer messaging for AI agent teams. No central orchestrator, no fixed topology. Spin up N agentic CLI sessions and let them self-coordinate over a shared SQLite bus.

View Source

tau

AM Core
Zig AI CLI

Non-interactive Zig reimplementation of pi. Single-shot AI CLI with tool-calling, goal mode, and sessions — designed for agents with JSON output and semantic exit codes.

View Source

sick-memory

Memory System

File-based memory system for AI coding agents with centralized storage, git-based scoping, and worktree support. Persistent knowledge storage across sessions using markdown files with YAML frontmatter.

View Source

goal-mode

Session Goals

Generic /goal command for Devin CLI, OpenCode, and Pi. Provides persistent session-scoped objectives that persist across turns with automatic continuation hooks.

View Source

DevOps Utilities

Lightweight utilities for daily workflows.

Commiat

CLI

Tired of writing git commit messages? Let AI do it for you! ✨

View Source

Adminer Node

UI

Lightweight database management UI for MySQL, MariaDB, PostgreSQL and SQLite.

View Source

hotify-cli

DevOps

Traefik/Cloudflare app management CLI with agent-friendly JSON output. Manage domains, SSL certificates, and Docker containers from the command line.

View Source

remotecmd-cli

Remote Execution

Execute commands on any machine, anywhere. WebSocket relay + token auth. Works over Tailscale, public internet, NAT, and firewalls with JSON output.

View Source

perrus-cli

Health Monitoring

Clean-room Go reimplementation of Gatus (10k+ stars) — agent-first health monitoring CLI. Track service uptime, response times, and system health with JSON output.

View Source

cmdcenter

Command Dashboard

Generic command execution dashboard with configurable commands via JSON config file. Run and monitor custom commands from a web interface with real-time output.

View Source

minipostiz-cli

Social Media

Offline social media scheduler CLI — agent-first, SQLite-backed, supercli plugin. Schedule posts across platforms without cloud dependencies.

View Source

easyredmine-cli

Project Management

Redmine API client for EasyRedmine. Read issues, post comments, edit descriptions. Designed for agents and automation workflows.

View Source

The Builder's Philosophy

Human expertise when you need it. Machine automation at scale. Built on open-source foundations.

Agent-First Design

Every tool outputs structured JSON, accepts deterministic arguments, and requires no TTY. Built to be consumed by Claude Code, Cursor, OpenCode, and any AI agent — not just humans.

Surgical Precision

One tool, one job. No daemons, no dashboards, no config sprawl. Every tool solves a concrete ops problem and fits in a 5-minute curl-pipe-sh install.

Production-Active

Not theory — every tool runs in production today, maintained by a single architect who ships code daily. Patterns that have survived real data, real traffic, real edge cases.

Open Core

90% open source — forever free. The 10% managed services (hosted relays, SaaS panels) exist for teams that don't want to operate infrastructure. No lock-in, no bait and switch.

Core Technologies

Language-agnostic foundations. Every solution ships as a single binary or Docker Compose recipe — zero dependency hell, zero vendor lock-in.

NodeJS
Runtime
EJS
Templating
Mongo
Database
Vue 3
Frontend
Tailwind
CSS Framework
Docker
Containerization
Traefik
Gateway
Go
Binaries
MFL
Your lang

Every solution ships as a single-file binary (Go, Rust, Zig, MFL) or Docker Compose recipe — no interpreters, no runtimes, no dependency hell. CLI+UI architecture designed for 2026 agent-native workflows. Stack-agnostic by design: the right language for the right job.

Machine-First Language

MFL — machin

Intrane bridges human and machine. MFL goes one step further — a language built for machines. Zero type annotations, one canonical function per line, compiled to native speed through C. No interpreters, no runtime, no human ergonomics tax.

machin is written in machin. The compiler is mature enough to compile its own source code back into itself — byte-for-byte identical, and as fast as before. A language that can rebuild itself from scratch is the line between a real, production-grade tool and a prototype. That's the kind of ground-up, verifiable engineering I bring, with AI, to the systems I build for clients.

0.18s
fib(40) — native C speed
✓
written in itself
89
releases shipped (v0.85)
MIT
Open source

Craft native-speed products in every domain

🌐
Web
machweb framework + WASM
⚙️
Backend
HTTP, SQLite, crypto, goroutines
🖥️
Desktop
C FFI → raylib GUIs, native TUI
📱
Mobile
Coming soon
func fib(n) { if n < 2 { return n } return fib(n-1) + fib(n-2) }

func main() { println(fib(40)) }

Canonical MFL — one function per line, types inferred, compiles to native

Explore machin on GitHub

v0.53 · MIT · Go-powered compiler, C-speed native output

Frequently Asked

Quick answers to the questions that usually come up.

What if I need more than 15 hours per month?
You can add extra hours at the Pay-as-you-go rate (€50/h). Or upgrade from AM Max to AM Enterprise for more included hours, dedicated Slack, and monthly architecture reviews.
Can I start with Pay-as-you-go and move to a bundle later?
Yes. Any Pay-as-you-go hours are credited if you subscribe to a bundle within 30 days. The bundle hours renew monthly — unused Pay-as-you-go hours don't roll over.
Do you work with my existing stack?
I work with any stack. AutoMaintainer runs on your own provider keys and integrates with GitHub, GitLab, and any CI system. Onboarding covers your conventions, playbooks, and no-touch zones.
What happens during onboarding?
I learn your infrastructure, configure AutoMaintainer workers, set up repo conventions, define playbooks, and mark no-touch zones. You review and approve before AM runs. Typical onboarding takes 2–4 hours.

How It Works

Subscribe, onboard, and let the Human + Machine loop run.

1

Pick a Bundle

Choose AM Max or AM Enterprise. Both include the full AutoMaintainer platform plus dedicated human support hours.

2

Onboard Your Stack

I learn your infrastructure, configure AutoMaintainer, set up conventions, and define your playbooks. Your repos, your keys.

3

24/7 Loop

AM runs automation around the clock. When you need architecture depth, integration, or incident response — I'm a message away.

One Invoice

Human support and AI automation bundled together. One subscription, one relationship.

Architecture Depth

A Senior Architect who knows your stack — not a ticket queue. Integration, incident response, and reviews included.

Built on Open-Source

AutoMaintainer is built on a2a-skill, SuperCLI, and 20+ OSS repos. Your stack, no lock-in.

Pricing

Human + Machine

Every plan pairs Human staff augmentation — a CTO/Architect you can lean on for AutoMaintainer, the Open Stack, or your own infra — with an AutoMaintainer service plan for AI-powered automation at scale.

→ Just want AI automation without the human layer? Explore standalone AutoMaintainer plans from €0/mo.

Pay-as-you-go

For one-off architecture reviews, integration work, or incident response.

€50 / hour

Billed per hour of work

  • Minimum 30 minutes per session
  • Payment due upon invoice
  • No recurring commitment
  • Architecture & integration work
  • No guaranteed response time
Inquiry

AM Max

AutoMaintainer Max + 5h human support

AutoMaintainer
€249 / month

Billed monthly

  • Platform
  • Includes AutoMaintainer Max managed infrastructure
  • Human Support
  • 5 hours included per month for support/integration
  • SLA response within 24h (business days)
  • Priority support queue
  • Cancel anytime

* Unused hours do not roll over

Inquiry

AM Enterprise

AutoMaintainer Enterprise + 15h human support

AutoMaintainer
€749 / month

Billed monthly

  • Platform
  • Full AutoMaintainer infrastructure management
  • Human Support
  • 15 hours included per month for support/integration
  • SLA response within 12h (business days)
  • Dedicated Slack / email channel
  • Monthly architecture review
  • Cancel anytime

* Unused hours carry over up to 1 month

Inquiry

Start with AI Staff Augmentation only — add human depth when you need it.

Explore AutoMaintainer Plans

Not ready for a bundle?

Start with AutoMaintainer Free — zero commitment, your keys, no markup. Upgrade to Max or Enterprise when your stack grows. Need a human architect? The bundles will be here.

Try AutoMaintainer free

Using any of these tools?

I'd love to hear how it's working for your team. Your story helps others decide.

Submit your testimonial
'