Labs
Book a callBook

AI that earns its place.Software that holds up.

Bx Labs is a web, mobile and AI agency, and a product studio, based in Paris. We build your website and your search visibility, your iOS and Android apps, your AI agents and your business software, your AI-made video.

One senior technical contact, from scoping through to production.

Tell us about your projectSee the services

Four services that combine on the same project

  • Website and SEO

    See the service

    Company site, landing page or shop: pages are rendered statically wherever the content allows it, so there is nothing left to compute when someone visits. Technical SEO, local SEO and your domain’s email are part of the work, not an option sold later. No ranking is promised — the site itself is sound and measurable.

    • A fast site, statically rendered wherever the content allows it
    • Technical SEO: URLs, markup, structured data, sitemap
    • Local SEO and Google Business Profile
    • Visibility in AI answer engines (GEO)
    • Business email on your own domain, with SPF, DKIM and DMARC

  • Mobile app

    See the service

    One codebase drives both iOS and Android, which cuts the work without taking anything away from what the user sees. Shipping to both stores is part of the engagement, with developer accounts opened in your company’s name. And because an app does not stop on release day, we take on what comes after: fixes, updates and OS version upgrades.

    • iOS and Android from a single codebase
    • Published on the App Store and Google Play
    • Push notifications
    • Offline mode, syncing once the network returns
    • A dedicated back end and API

  • AI agents and business software

    See the service

    An agent that answers from your own documents, an automation that absorbs a process costing hours every week, an internal tool that replaces a spreadsheet nobody can hold any more: the starting point is always a specific task, never a demo. The tool plugs into the systems you already run rather than forcing you to key everything in twice. Once in production, it stays supervised.

    • Agents wired to your real data
    • Automating the processes that cost hours every week
    • Business applications and internal SaaS
    • CRM, ERP and invoicing integrations
    • Supervision in production: logs, alerts, fixes

  • AI video production

    See the service

    Short videos made with generative models: advertising, product demos, recurring social content. The model speeds up the making; it decides nothing — the brief, the edit and the sign-off stay human. What AI genuinely changes is the cost of the variant: once the format holds, the next series no longer costs a shoot.

    • Vertical and horizontal formats, cut per network
    • Synthetic voice, subtitles and multilingual versions
    • Avatars and presenters, with the rights in writing
    • Editing, grading and artefact correction
    • A production chain you keep: run the next series without us

Four stages, no six-month tunnel

  1. Scope

    A call, your tools, what is costing you time. We name the business problem, the number that will say whether it worked, and the first piece of work to open.

  2. Design

    Flows, mock-ups, architecture and model choices. You see what it will look like and what it is built on before a single line is written.

  3. Build

    Short cycles, a demo at the end of each one, and direct access to the person writing the code. No six-month tunnel, no account manager in between.

  4. Operate

    Release, logs, fixes, scaling: the system stays watched after it goes live. And the code is handed over to you.

An independent studio, run hands-on

Our toolchain

TypeScript, JavaScript, Node.js, Python, PostgreSQL, HTML5, CSS, Bash, React, Next.js, Vite, TanStack, Tailwind CSS, CSS Modules, shadcn/ui, Radix UI, Lucide, React Hook Form, Zod, i18next, Three.js, Webflow, Fastify, Express, NestJS, Convex, FastAPI, Flask, Streamlit, WebRTC, Model Context Protocol, Anthropic, Claude, Supabase, Neon, Firebase, SQLite, Drizzle, Prisma, SQLAlchemy, Better Auth, Google OAuth, JSON Web Tokens, Bitwarden, Google Cloud, Google Pub/Sub, Docker, Cloudflare, ngrok, Aircall, Intercom, Meta, WhatsApp, Instagram, Messenger, Telegram, Brevo, Resend, Stripe, Cal.com, Notion, Sentry, Vitest, Jest, Testing Library, ESLint, Prettier, Biome, Git, GitHub, GitHub Actions, pnpm, npm, Turborepo, FFmpeg, pandas, NumPy, scikit-learn, Plotly

No account manager in between: you talk directly to the senior technical founder who runs the studio — the person who designs, writes the code and ships it. We deliberately cap how many engagements run in parallel — that cap is what makes the attention on each of them real.

The studio also publishes software of its own: ezjob and 4bl1ty are in production, agent-secrets is open source, and Olvia — our conversational agent — answers your customers on your website, on your messaging channels and on your inbound calls. The catalogue lists each one with its real status.

See the products

Frequently asked questions

Who owns the code?

You do, all of it. The source code, the repository, the domain name and the hosting accounts are created in your name or transferred to you at the end of the engagement. We keep no access you would depend on: changing supplier must never cost you your tool.

How long before a first version is in production?

A few weeks, provided the first scope is cut down to what actually solves the problem. It depends on the size of the project, on whether your data is available, and on the sign-offs you need on your side; we do not give you a date before scoping, and a date given without scoping commits nobody.

What kind of companies do you work with?

Independents, small and mid-sized companies with no in-house IT, and product teams that want senior technical reinforcement on one specific piece of work. The number of engagements running in parallel is deliberately capped: that is what keeps us reachable and keeps the cycles on time.

What does a first conversation look like?

A call to understand your business, the tools you use today and what is costing you time. You then receive a written scope — the problem as we understood it, what is feasible, what is not — followed by a proposal with a first scope and a plan. Nothing is committed until that document suits you.

What happens after go-live?

Going live is not the end of the engagement. Logs and errors are watched, fixes are applied, and scaling is dealt with as usage grows. You know who to write to, and it is the same person as during development.

How is data handled when a language model is involved?

That is an architecture decision, and it is settled during scoping: a call to a model API, execution on infrastructure you control, filtering or anonymising before anything is sent. We tell you what leaves your systems, where it goes and what is retained, according to how sensitive the data is. We sell no certification and no promise of compliance.

Tell us what you want to automate, or to build

Book a 30-minute call with a senior technical contact: the context, the tool involved, the deadline. Pick your slot in the calendar, without leaving this page.

Prefer to write?Paris — GMT+1 · Reply within 24 working hours

Book a 30-minute call

Allow Cal.com to view available times and book here.

The calendar is provided by Cal.com. It receives the technical data of your connection, sets a cookie on its own domain, and it is the one that handles the booking.

What that involves is set out in our privacy policy.

Calendar not showing? You can still book on cal.com.