Source: Conversations & Memory overview (Microsoft Agent Framework docs)
AgentSession usage in Agent Framework: create, reuse, serialize and resume sessions, local vs service-managed history, and the tenant-isolation caveat for service conversation IDs.
Contents
- What: official documentation page, Microsoft Learn
- Author: Microsoft (eavanvalkenburg)
- Date: ms.date 2026-05-28 (read 2026-11-14)
- Original: https://learn.microsoft.com/en-us/agent-framework/concepts/agents/conversations/
Summary
AgentSession carries conversation context between runs. Sessions are either local (history stored by a history provider) or service-managed (an opaque server-side conversation ID). The page gives the create / pass / persist / resume pattern for C#, Python and Go and points to sub-pages on context providers, compaction, storage and the chat-history memory provider.
Key claims
- Pattern: create a session, pass it to every run, persist it (
SerializeSession/session.to_dict()), and rehydrate (DeserializeSessionAsync/AgentSession.from_dict) or resume by service conversation ID (get_session(...)). - OpenAI Responses and Conversations IDs are scoped to the API key or project; if one hosted agent serves many users, store IDs server-side and check the authenticated user or tenant before resuming.
- Context providers, compaction and storage are separate, pluggable concerns.