image_How-it-works

How it works

From setup to a working platform in three steps.

Step 1Auto-Processing

1. Create a workspace and upload your content

A workspace is your project, your client, or your department. Inside, create folders for each workstream — training-videos, client-deliverables, project-website, process-models.

Upload files and they're processed automatically. Videos become streaming-ready players. Images are optimized to WebP with thumbnails. Documents render inline.

Step 2Folder Apps

2. Turn folders into purpose-built apps

Assign a type to any folder and it becomes a specialized application:

Media gallery — Searchable archive with streaming playback and tagging.

Training course — Structured learning environment with presentations and video modules.

Process modeler — Interactive BPMN editor and simulator.

Website — Multi-language site builder with AI-assisted content.

No configuration needed. The folder is the application.

Step 3AI AgentsYour Infrastructure

3. Work together with AI agents

Open any folder and compatible AI agents appear. A content writer for your website. A course planner for training materials. A process analyst for your BPMN diagrams.

Give an instruction or pick a quick action. The agent reads your files, understands the structure, and creates or refines content. You approve, adjust, or let it run autonomously.

Your API keys. Your choice of model — Anthropic, OpenAI, or local via Ollama. Everything stays on your server.

Access CodesExpiry & Revoke

Share with anyone — no accounts required

Generate an access code for any folder. Send the link. Your audience opens it and sees everything — videos, documents, training courses, process models.

No sign-up forms. No password resets. Create different codes per audience. Set expiry dates. Revoke when the project ends.

HostedFederationStandalone Delivery

Three ways to deliver when you're done

Keep it on your platform: Clients access via codes on your infrastructure. You control the environment, they consume the content.

Federate across servers: Connect multiple Aipokit instances to share catalogs across your organization or with partners.

Hand over standalone: Export a standalone binary for the client to run on their own server — their data, their branding, their infrastructure.

What people use it for

Two examples from real workflows.

EducationAccess Codes

Running a workshop

You prepare a training session. Your AI agent drafts the agenda from your notes. Presentations run directly from the platform. Session recordings upload and auto-transcode to streaming afterward.

Share one access code with all participants. They revisit the materials weeks later — no accounts needed.

ConsultingData Sovereignty

Delivering a client project

The workspace holds process models, interview recordings, training materials, and a project website — all in one place. AI agents analyze processes and draft recommendations.

When the client asks where their data lives, you point to their server. No third-party routing.

Early Access

Questions? Let's talk.

Aipokit is in active development. Early access for businesses, educators, consultants, and teams who want a platform they own.

Get in touch via aipokit.com or reach out on LinkedIn.