Built in India · Serving India, Middle East, MENA, West Asia & Africa EST. 2015 · 11 YEARS
Code Folks Technologies
AI · Service-based development

Rumi_ walks the repetitive road, so your team can soar above it.

Rumi_ is not a product you configure. It is an AI engagement: our engineers study how your business actually decides, then build agents that carry the repetitive work end to end — inside your systems, under your rules, with your people still in charge of the judgement calls.

Built on large language models, the Model Context Protocol, n8n automation, ElevenLabs voice and retrieval over your own documents.

Rumi_ AI agent
Rumi_ · on duty
{{ s.v }}
{{ s.l }}
The engagement

How a requirement becomes a working agent

Seven stages, each with something you can inspect at the end of it. Nothing goes autonomous before it has run beside a human and agreed with them.

{{ s.n }}
{{ s.phase }}

{{ s.t }}

{{ s.d }}

{{ s.out }}
Architecture

What sits between a request and an action

Five layers. A request enters through any channel, is reasoned about with your own context, and leaves as a write into a system that already holds your data.

{{ l.n }}
{{ l.t }}

{{ l.d }}

{{ p.t }}
The stack

What Rumi_ is built with

Each piece is chosen for a job, not for a logo. Anything can be swapped for a self-hosted equivalent if your data policy requires it.

{{ s.k }} {{ s.t }}

{{ s.d }}

Used for: {{ s.use }}
Automating the know-how

How the knowledge in your team's head becomes a rule the system runs

The hardest asset to automate is not the data — it is the judgement that lives with the person who has done the job for nine years. Rumi_ moves it out of their head in four deliberate steps, with them in the room.

{{ m.n }}
{{ m.t }}

{{ m.d }}
What you get

Benefits, stated plainly

No claim here needs a leap of faith. Each one is measured against your own baseline before and after the agent goes live.

{{ b.t }}
{{ b.d }}
Guardrails

Where the agent stops and a person decides

{{ g.t }}
{{ g.d }}
Where it earns its keep

Work Rumi_ is already shaped for

{{ c.k }}
{{ c.t }}

{{ c.d }}

Start with one process, not a transformation programme.

A discovery call, one shadowed process and a pilot agent. If it does not beat the manual baseline, you have lost a fortnight rather than a budget.

Book a discovery call