Integrated open-multi-agent with TencentDB-Agent-Memory via the Hermes Gateway

[JackChen-me]

We just added a cookbook to open-multi-agent (a TS multi-agent orchestration framework) that uses TDAM as the long-term memory backend for an agent team, through your Hermes Gateway: examples/integrations/with-tencentdb-memory. Team results capture into TDAM via /capture and come back on the next run via /recall, with the L0→L1 pipeline distilling in between.

Two things we hit, in case they help other integrations: the L1 extractor is user-centric (it distills persona/episodic/instruction memories about the user and skips assistant output), so the agent's result has to go in the user slot of the captured turn or nothing gets extracted; and POST /session/end drains in-flight extraction but doesn't force turns still under the everyNConversations threshold, so we run the Gateway with everyNConversations: 1 for deterministic capture-then-search in a short demo.