WithCompactionModel sets a dedicated model for session compaction (summary generation), letting a heavyweight primary model delegate the slow, expensive whole-conversation summary call to a smaller/faster one.
(model provider.Provider)
| 95 | // generation), letting a heavyweight primary model delegate the slow, expensive |
| 96 | // whole-conversation summary call to a smaller/faster one. |
| 97 | func WithCompactionModel(model provider.Provider) Opt { |
| 98 | return func(a *Agent) { |
| 99 | a.compactionModel = model |
| 100 | } |
| 101 | } |
| 102 | |
| 103 | func WithSubAgents(subAgents ...*Agent) Opt { |
| 104 | return func(a *Agent) { |
no outgoing calls