AI assistant
Connecting an AI provider
Siteho's free daily requests, connecting your own OpenAI-compatible provider with its address, key and model, what each costs, and the messages you may see when a provider refuses.
Siteho's free requests
Until a workspace adds a provider of its own, Siteho gives it some free requests a day for the assistant (Siteho's free key). The AI Assistant page says how many and how many are used today; when they run out the assistant stops until tomorrow (UTC) or until you add a provider.
In the app: Workspace › AI Assistant › Setup
Connecting your own provider
- Open the workspace's AI Assistant › Setup.
- Pick a provider chip. It fills in the API address; Create a … API key opens the provider's page.
- Paste your API key (OpenAI-compatible) and type the Model exactly as the provider names it. It must support tool calling.
- Press Save AI Settings. Requests now go to your model: the bar says Workspace's own key.
| Provider | API address | Example model |
|---|---|---|
| OpenAI | https://api.openai.com/v1 | gpt-4.1-mini |
| OpenRouter | https://openrouter.ai/api/v1 | openai/gpt-4.1-mini (one key for many makers; some free) |
| Google Gemini | https://generativelanguage.googleapis.com/v1beta/openai | gemini-2.5-flash |
| Anthropic | https://api.anthropic.com/v1 | claude-sonnet-4-5 |
| Groq | https://api.groq.com/openai/v1 | llama-3.3-70b-versatile |
| DeepSeek | https://api.deepseek.com/v1 | deepseek-chat |
| Mistral | https://api.mistral.ai/v1 | mistral-large-latest |
| Together AI | https://api.together.xyz/v1 | meta-llama/Llama-3.3-70B-Instruct-Turbo |
| AIHubMix | https://aihubmix.com/v1 | gpt-4.1-free (free models with a daily allowance) |
| Other | Any https address: the part before /chat/completions | As your provider writes it |
Your key
- Stored encrypted
- Never shown again or sent to the browser; only used for this workspace's requests.
- Tied to its address
- A new API address needs the key typed again: the saved key is only ever sent to the address it was saved for.
- Keeping it
- Leave the key field empty to keep the saved key.
- Remove Key
- The assistant stops using your provider; meanwhile it uses Siteho's free requests, if any.
- Costs
- Siteho doesn't charge for the assistant: your provider bills its requests at its prices.
When a provider refuses
| Message | What to do |
|---|---|
| The AI provider rejected the API key saved for this account. | Check the key in your provider's dashboard and save it again. |
| The AI provider's per-minute limit was reached. | The assistant waits and tries again by itself, three times. |
| Your AI provider says today's limit for this key or model is reached. | Wait for the provider to reset it, or pick another model. |
| The AI provider has no model named … | Check the Model and the API address. |
| Your AI provider wants your account topped up… | Add credit at your provider: its own words are shown. |
| Today's free requests are used up… | Wait for tomorrow, or connect your own provider. |
No key at all
Use any AI chat by copy and paste instead, at no cost and with none of these requests: see AI without a key.