Getting started
The platform has two services. Which one you use depends on the capability — each capability page tells you the host and key to use.
| Service | Base host | Auth header |
|---|---|---|
| Gateway — chat, RAG, translation, vision, standard speech-to-text, text-to-speech | https://ai.ollalink.com |
Authorization: Bearer <key> |
| Speech GPU — high-accuracy & realtime speech-to-text, 50-voice TTS, voice cloning, voice design, live translation | gpu-*.ollalink.com |
X-NH-GPU-Key: <key> |
Full detail and error codes: Authentication.
Your first request (Gateway)
Send your partner key as a Bearer token. Keep it server-side — never ship it to a browser.
bash — curl
curl https://ai.ollalink.com/v1/chat/completions \
-H "Authorization: Bearer sk-partner-XXXXXXXXXXXX" \
-H "Content-Type: application/json" \
-d '{
"model": "chat",
"messages": [
{"role": "system", "content": "You are a helpful networking assistant."},
{"role": "user", "content": "In one sentence, what does BGP do?"}
],
"max_tokens": 200
}'
Python (OpenAI SDK)
python
from openai import OpenAI
client = OpenAI(base_url="https://ai.ollalink.com/v1",
api_key="sk-partner-XXXXXXXXXXXX")
resp = client.chat.completions.create(
model="chat",
messages=[{"role": "user", "content": "Explain OSPF in two sentences."}],
max_tokens=300,
)
print(resp.choices[0].message.content)
Model lanes at a glance
| Model | Use it for | Languages | Tokens |
|---|---|---|---|
chat |
General chat, RAG, summarization, vision, tool calling | Multilingual | Large context; size max_tokens to the task |
code |
Dedicated coding — reasons first | Code + natural language | max_tokens ≥ 4000 (ideally 8000), timeout ≥ 120 s |
local |
Fully private, on-hardware chat — prompts/outputs never leave our infrastructure | Primarily English | Keep max_tokens modest; prefer stream: true |
translate |
Technical translation — keeps IPs, acronyms, CLI in Latin | Broad (Indian + European) | Short outputs |
Call GET /v1/models for the live list.
Health check
bash
curl https://ai.ollalink.com/healthz # -> ok
Where to next
- Chat & RAG — chat, RAG, coding, vision, tools (languages + token limits per lane)
- Translation
- Speech-to-text — standard, high-accuracy, realtime
- Text-to-speech — gateway voices + the 50-voice catalog
- Voice cloning
- Live translation