OpenAI — Integration

Overview

Use OpenAI models for chat, tools, and multimodal prompts. Configure credentials in the KeyVault and call via Query API or the internal handler.

Capabilities

• Text & multimodal prompts (images via data URLs)

• Provider-aware usage tracking & per-day rollups

• Optional moderation check before/after generation

• Fallback routing to other providers on transient failures

• Memo-cache for deterministic prompts (24h)

1. 1

   Step 1

   Go to AI Hub → Provider Keys or POST /AdminApi.php with {"provider":"openai","creds":{"api_key":"sk-…"}}.

   Back to top
2. 2

   Step 2

   Test with POST /QueryApi.php {"provider":"openai","model":"gpt-4o-mini","prompt":"Hello"} and verify {"ok":true}.

   Back to top
3. 3

   Step 3

   Enable moderation mode (block/warn/allow) in policy settings if desired.

   Back to top
4. 4

   Step 4

   Optionally enable Redis for memo-cache to reduce latency and cost.

   Back to top

Limitations

• Subject to OpenAI rate limits and content policies.

• Image/audio support depends on model capabilities configured at call time.

• Costs accrue per token; ensure quotas are set to prevent overruns.

FAQs

Pricing