Talk to the engine
SAPIR 861 · ENGINE
Talk to the engine
What answers here is not a language model but the engine itself. Every line of the answer is the result of a call you can repeat by URL. That is why the answers never vary and contain no invention.
How it works
- The phrase is parsed into an intent and data — by rules, with no model.
- The intent calls one of our open tools: counting, the attestation base, the coincidence detector.
- The answer is assembled only from what the call returned. Empty is said as empty.
- The thread is kept: “and Sarah?” or “who recorded it?” refer to the previous answer.
The same conversation runs in Telegram: SAPIR 861 · MCP · /api/talk
Open Scriptures Hebrew Bible (WLC), CC BY 4.0