MICROSIVE

The tools we build with

The software, AI models and voice tools behind our own products and the work we build for clients, grouped by the job they do.

Reviewed by: Rahul Jaiswar, founder

These are the tools we use or have tested in our own products and builds. A tool your business relies on may not be listed here, and that doesn’t rule it out (tell us what you run on).

Web and app front end

  • TypeScript: JavaScript with types, so mistakes surface before release.
  • Next.js: public pages and logged-in screens in one codebase.
  • React: screens with filters, selections and multi-step forms.

Server and background work

  • Node.js: APIs, webhooks and integrations between systems.
  • Python and FastAPI: data work, scoring, audio, and jobs on a schedule.

Data

  • PostgreSQL: business records, with each client’s data kept apart.

Hosting and deployment

  • Docker: an app and its setup packaged to run the same everywhere.
  • Railway: hosting for apps with workers and a database, in your account.

AI models

  • OpenAI: models we have tried in voice agent experiments.
  • Anthropic: Claude in internal tooling, and the Claude connector in Microsive Dialer.

Voice and calling

  • LiveKit: self-hosted real-time voice for our voice agents.
  • Telnyx: US numbers for Microsive Dialer, with recording and cost per call.

Where these tools end up

What we build with these tools is usually the same shape: capture what happened where the work happens, tie it to the order, load or client record it belongs to, and keep it beside the system you already run. Warehousing and 3PL and trucking set out what that looks like problem by problem, and how we work covers scope, quotes and handover.

Already have a stack? We can start from it.

List the systems your business runs on and what you want built next. Someone from our team replies within one business day.