Provider guide

Starting with
Hrida.ai

Use the Hrida.ai model API as your provider: curated models served from api.hrida.ai, with no GPUs or model downloads on your side. It speaks the OpenAI Chat Completions protocol, so it connects like any OpenAI-compatible provider.

Request an API keyOther providers

OpenAI-compatibleManaged by Zlabs InnovationNo model downloads

Connection details

Setting	Value
Protocol	OpenAI-compatible (Chat Completions)
Base URL	https://api.hrida.ai/hridaai/v1
API key	Your Hrida.ai API key
Models	Auto-detected from the /models list

1. Get your API key

Hrida.ai API keys are issued by Zlabs Innovation. Request one through the contact page, or ask your Zlabs account contact if your organization already has access. Keep the key secret — anyone with it can make requests billed to your account.

2. Add the connection

1. Open Admin Settings → Connections.
2. Under OpenAI API, click Manage (the wrench icon), then Add New Connection.
3. Fill in:
   • Connection Type: External
   • URL: https://api.hrida.ai/hridaai/v1
   • API Key: your Hrida.ai API key
4. Click Save.

Leave Model IDs empty to show every model your key can use, or list specific IDs to limit what users see. If you also connect another provider with overlapping model names, set a Prefix ID such as hrida/ to tell them apart.

3. Verify the key (optional)

Check the key and see which models it unlocks before adding it:

curl https://api.hrida.ai/hridaai/v1/models \
  -H "Authorization: Bearer YOUR_HRIDA_API_KEY"

A JSON list of models means the key works. {"error":"Unauthorized"} means the key is missing, mistyped, or not yet active.

4. Start chatting

Pick a Hrida.ai model from the model selector at the top of a new chat. There is nothing to download — requests go straight to the hosted API.

Self-hosting instead?

To run models on your own hardware, see LocalAI or any server listed in OpenAI-Compatible. You can connect Hrida.ai alongside them and switch per chat.

Troubleshooting

• No models appear after saving. Re-check the URL ends in /hridaai/v1 with no trailing slash, then run the curl check above.
• Unauthorized (401). The key is wrong, revoked, or not yet active — request a new one.
• Model list loads slowly or times out. See Model List Loading Issues.