What exactly does Oktya do?

We build three things for your business: AI agents that handle, qualify and sell on WhatsApp and Instagram 24/7; custom software (management systems, internal panels, e-commerce, premium sites); and Oktya Studio, the panel where you watch and control your agent live.

How is this different from someone just installing a bot?

Oktya Studio. Anyone can leave a bot running blind. We give you the panel to see every conversation, real-time KPIs, and a chat to talk to your own agent. You see what it does and you steer it.

Do I need a technical team?

No. We set up and maintain everything. The agent works on top of the tools you already use (WhatsApp, Instagram, your CRM) and you control it from Studio, without writing a single line of code.

Which channels does the agent work on?

WhatsApp and Instagram, 24/7. It replies in seconds, qualifies leads, books, follows up, and escalates to a human when needed.

How fast can you get it running?

It depends on scope, but an AI agent in production is usually live in days, not months. On the first call we give you a concrete timeline.

Does it integrate with my CRM and payment rails?

Yes. We connect the agent to your CRM, e-commerce and payment rails so it qualifies, charges and logs everything where you already work. Custom integrations for your stack.

What languages does it support?

Spanish and English. The agent respects the language of each conversation.

How does billing work?

A monthly payment per business, scaled to scope: which agents, which software and which integrations we build. We tell you clearly on the first call, no surprises and no long contracts.

Do you work with businesses outside my country?

Yes. We work with SMBs across Latin America, fully remote, in Spanish and English.

LIBRARY

Infinite b-roll of yourself with AI

Record a few real clips one afternoon. The AI multiplies them into different versions — different pose, different outfit, different angle — giving you months of content without recording again.

afternoon of recording

∞

generated clips

~26

credits per clip

Hector Mejia · @hectormej.ia

THE PROBLEM

B-roll of yourself is the hardest to get

You can generate landscapes, mockups, abstract visuals without leaving your desk. But clips of you in action — from behind, from the side, hands working — require a camera, lighting, setup, time.

And when you record them, it's always the same angles. Always the same outfit.

THE SOLUTION

You record once with good lighting. The AI reimagines those clips into variants you could never record alone.

THE PIPELINE

Three steps. One afternoon of recording.

PASO 1

Record real clips

Fixed camera, daylight, different angles: from behind, profile, hands on the keyboard. These are your raw material.

PASO 2

Nano Banana 2 reimagines each frame

The real frame comes in as an identity reference. The AI generates a new image of you doing a different action, in different clothes, from a different angle. Same person. Completely different scene.

nano_banana_pro · 2k resolution

PASO 3

Seedance 20 animates it.

The generated image enters Seedance as a starting point. Out comes a 5-second clip, fixed camera, with you moving inside the scene.

seedance_2_0 · 5s

HARD RULE

Never animate a frame directly from the video. First Nano Banana reimagines it with a different action, then that gets animated. If you skip this step, you inherit the original pose from the footage — talking head, mic in hand, mid-gesture.

THE COST

What each clip costs

Everything runs on Higgsfield. Credits are consumed like this:

Image (Nano Banana 2)~2–4 credits

Video 5s (Seedance 2.0)~22 credits

Total per full clip~26 credits

TO GET STARTED

With Higgsfield's Plus plan (~$39/mo) you have enough credits to generate dozens of clips per month. Start with 3-5 test clips to calibrate the style before doing a large batch.

THE APP

Higgsfield. Just Higgsfield.

Two models inside the same platform:

Nano Banana 2

Image generation. Pass the real frame as a reference, describe the new scene, and it generates you in a different situation.

~2–4 credits / image

Seedance 2.0 (image-to-video)

Image animation. The image generated by Nano Banana goes in here and out comes a 5-second video.

~22 credits / 5s video

NEXT LEVEL

Prefer doing it from Claude?

Higgsfield has an MCP server you can connect to Claude Desktop or Claude Code. Instead of opening the Higgsfield web UI, you describe what you want directly to Claude — and Claude generates the images and videos for you. Same pipeline, operated from chat.

Search for 'Higgsfield MCP' in the official docs or in Claude Desktop → Settings → MCP Servers. Connect once, use forever.

TECHNICAL DETAIL

Always request 'nano_banana_pro' with 2k resolution. There's an internal routing where the 'nano_banana_2' model can serve a flash version that doesn't maintain your identity. The pro version gives you the real model.

THE RULES

What I learned the hard way

Fixed camera always. You provide the movement, not the camera. If the prompt says 'handheld' or 'kinetic', the result looks like it was shot by a drunk drone.
Frame as reference, not as seed. The real frame tells the AI who you are and what your space looks like. It doesn't tell it what to do. You design the new scene from scratch.
Daylight for recording. If you record at night with red LED backlight and generate daytime, the clips won't match. Daylight = the easiest set for AI to replicate.
Face-on b-roll comes from real footage. For clips where your face is visible from the front, better use the original footage. The AI maintains identity in 85% of cases — good enough for back or profile shots, not for close-ups.

Want to implement something like this for your business?

Oktya builds AI agents and automation systems for LATAM businesses.

Talk to Oktya

Oktya · 2026