Memory · Software component
Episode Summarizer
Software componentMemoryCognition & Memoryarc:EpisodeSummarizer
A consolidation component that prompts a language model to compress a verbose episode into an essential narrative of problem, root cause, solution, outcome metrics and lessons learned.
Responsibility. Compresses individual raw episodes into compact decision-relevant summaries.
Also known as: Episode compressor, Lossy consolidation, Session summaries
Relationships
invokes dependency
is invoked by dependency
writes dependency
is triggered by dynamic
Design guidance
- SHOULD preserve problem, root cause, solution steps, outcome metrics and lessons while discarding phrasing and routine procedural detail.
- MAY retain original episodes temporarily before applying lossy summarization so recent cases keep full detail.
Quantitative guidance
As stated by the sources; verify before use.
- Summaries occupy 5-10% of original episode size while retaining 80-90% of decision-relevant information (Ch5.7).
- Summarization reduces storage by 90%+ (Ch5.7).
- Originals may be retained 30-90 days before lossy consolidation (Ch5.7).
Classification
- Patterns
- LLM-prompted structured summarization
- Quality attributes
- Performance efficiency (ISO/IEC 25010)
- Risks mitigated
- Prohibitive cost of storing full transcripts