OpenAI — Integration

Overview

Bring-your-own key model connector. Use OpenAI models for intent, summarization, tool-use, and reply drafting under Chat Core guardrails and budgets.

Capabilities

• Model allowlists & per-team budgets

• Prompt firewall and redaction policies

• Tool/function calling support with schema contracts

• Cost & latency metrics to InsightLake

1. 1

   Step 1

   In Settings → Models, add an OpenAI provider with your API key.

   Back to top
2. 2

   Step 2

   Select allowed models and set monthly budgets/alerts.

   Back to top
3. 3

   Step 3

   Attach the provider to a pipeline (e.g., intent → tools → reply).

   Back to top
4. 4

   Step 4

   Run evals with replay sets and compare outcomes.

   Back to top

Limitations

• Data residency depends on provider region; configure per-policy.

• Throughput subject to provider rate limits and quotas.

FAQs

Pricing