a request orchestration system for multiple LLM providers with a quorum mechanism and answer synthesis via a “judge”.
The quorum mechanism helps filter out erroneous or contradictory responses from individual LLMs. When the majority of models agree on something, the likelihood that the answer will be correct is significantly higher.
Fault tolerance (if one provider is unavailable, it does not participate in the quorum);
Flexibility (you can select providers based on tasks and budget);
Protection against vendor lock-in (you’re not tied to a single supplier).
The “Honest Status” feature is critical for building user trust. By knowing how many providers actually participated in generating the response, users can assess its reliability.
A dedicated “Judge” module solves the problem of simple averaging or voting. It can:
extract the best elements from different responses;