Senaiy Lab — API Overview

Senaiy Lab is a platform for building AI into your business without giving up control. You get three layers, all behind one base URL and one API key:

Models — an OpenAI-compatible endpoint for chat, streaming and embeddings. Point any OpenAI SDK at us and it works: GPT, Claude, Gemini and regional models through a single interface.
Agents — AI assistants you configure in the portal: their instructions, their knowledge, and exactly which tools they may touch. Sensitive actions pause for a human approval before anything is written to your systems.
Workflows — multi-step processes drawn as a graph: branch on results, gate on approvals, call several agents in sequence — durable across hours or days.

your code ──▶ https://platform.senaiy.ai/v1
                 ├── /chat/completions, /embeddings, /models   (OpenAI-compatible)
                 ├── /agents, /runs                            (governed assistants)
                 ├── /workflows, /workflow-runs                (durable processes)
                 ├── /collections, /files                      (your knowledge)
                 └── /usage                                    (every request, rated in USD)

Why teams choose Senaiy Lab

Nothing writes without permission. Every agent runs with an explicit tool allowlist; write actions can require a human "yes" first. Every step is audited with its cost.
One key, one wallet. All three layers share your API key and prepaid balance. No seats, no subscriptions — you pay for what runs, and auto-reload keeps work from stopping mid-task.
Arabic and English are equals. Models, portal and agents handle both natively.
Your data, your borders. Residency policies can restrict an organization to in-country engines; violations are refused with a clear error, not logged quietly.
Built to survive. A workflow waiting three days for an approval loses nothing — runs are checkpointed and resume where they stopped.

Where to go next

New here? Read Quickstart — you will make your first call in five minutes. Then: Authentication for keys and scopes, Inference for the OpenAI-compatible surface, Agents and Workflows for orchestration, Knowledge for RAG.