LlamaIndex Memory
Two-tier model: ChatMessage FIFO queue with overflow flushed into long-term Memory Block objects. Memory class replacing older ChatMemoryBuffer.
Active · Verified 2026-10-22 · fresh as of 2026-10-22 · 61 cited facts, 2 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Type
- FIFO short-term + Memory Blocks vendor fix
- Tier
- T2
- Section
- Framework-embedded memory
- Latest release
- v0.14.25 (2026-09-21) repo fix
- Pricing
- Free + paid vendor fix
- Funding
- $28M total Series A · 2025-03 vendor fix
- Freshness
- 2026-10-22 fix
- Layer
- agent-infra repo fix
- Accelerator
- aws-genai-accelerator-2025 third-party fix
- SDK languages
- python, typescript repo fix
- Developer activity
- OpenRank 26.99 (12-mo avg 28.66) · 261 participants — 2026-09 third-party fix
Memory taxonomy
- storage
- kv
- retrieval
- injection
- persistence
- cross-session
- update
- append-only
- unit
- episode
- governance
- inspectable
- conflict
- append-only
When to use
Optimised for: FIFO + Memory Blocks for chat agent context
Anti-fit: not for non-Python stacks
Pros & cons
Pros
Mature retrieval ecosystem behind it — memory benefits from years of LlamaIndex retrieval work.
Cons
Retrieval-tilted — memory feels like a RAG add-on rather than a first-class abstraction.
Claims & capabilities
May 2025 introduced flexible Memory API combining short-term FIFO chat history with pluggable Memory Blocks (StaticMemoryBlock, FactExtractionMemoryBlock, VectorMemoryBlock); 35% retrieval-accuracy improvement claim (overall framework, not memory-specific)
Technical surface
- API surface
- SDK: Python, TS
- Backend storage
- pluggable
- Deployment
- Both
- Embedding model
- multiple supported
- Multi-tenancy
- not applicable — library; namespace per index/session
- MCP
- via official adapter — llama-index-tools-mcp
- A2A
- supported — LlamaIndex A2A agent server
- OpenTelemetry
- via adapter — Arize Phoenix / OTel
Alternatives to LlamaIndex Memory → live options ranked by shared design, licence, hosting and MCP support
Similar systems
Other framework-embedded memory in the catalog, ranked by inbound references.
- Spring AI ChatMemory T3
Wraps conversation memory behind an Advisor abstraction. MessageWindowChatMemory default 20-msg sliding window (preserves system messages); pluggable ChatMemoryRepository backends (in-memory, JDBC, Cassandra, Neo4j). Memory injected as cross-cutting concern at the ChatClient level.
- Flowise Memory T1
Discontinued: Flowise (acquired by Workday, Aug 2025) was sunset in 2026 — code freeze 2026-07-29, GitHub repo archived 2026-08-13, end of life 2026-08-31. Was a visual LangChain canvas with LangChain-native memory nodes (Buffer, Buffer Window, Conversation Summary Buffer) backed by an internal chat_message DB table; Buffer Window kept the last K interactions, Summary Buffer compressed older turns while keeping recent ones verbatim.
- LangGraph Persistence T2
Distinct from LangMem. Built-in checkpointer saves graph state per super-step (short-term, thread-scoped). Store System adds long-term hierarchical key-value memory across threads with optional vector search + TTL. Postgres / Mongo / Redis stores for production.
- AutoGen Memory T2
AutoGen's memory layer: ListMemory chronological context plus teachable agents that vectorise corrections; integrates with Mem0/Zep rather than building deep memory natively. Microsoft AutoGen is in maintenance mode (community managed, bug and security fixes only), and Microsoft points new projects to Microsoft Agent Framework.
- CrewAI Memory T2
Memory subsystem inside the CrewAI orchestration framework; integrates with Mem0 for the long-term tier.
- Hermes Agent memory (Nous Research) T3
Nous Research's open-source self-improving agent keeps an agent-curated MEMORY.md/USER.md notebook injected at session start, full-text (FTS5) search over past sessions that returns raw messages without LLM summarisation, and optional Honcho user modelling. one of the most-starred agent projects of 2026.
Sources (60 cited facts, 2 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is FIFO short-term + Memory Blocks vendor
- What it does & what's distinct Two-tier model: ChatMessage FIFO queue with overflow flushed into long-term Memory Block… vendor
- Claims / benefits May 2025 introduced flexible Memory API combining short-term FIFO chat history with plug… vendor
- Modalities text repo
- Created 2022-11 vendor
- Latest release v0.14.25 (2026-09-21) repo
- License MIT repo
- GitHub run-llama/llama_index — ~52.6k stars, last commit 2026-09 or later repo
- Funding $28M total Series A · 2025-03 vendor
- Customers / scale 11-50 ppl vendor
- Pricing Free + paid vendor
- Compliance SOC 2 Type II vendor
- Deployment Both vendor
- HQ US (San Francisco) vendor
- Founders / pedigree LlamaIndex; co-founder Jerry Liu (ex-Uber ML; Carnegie Mellon grad); YC-backed vendor
- Memory volume / scale no scale claim repo
- API surface SDK: Python, TS vendor
- Backend storage pluggable vendor
- Embedding model multiple supported vendor
- Versioning immutable append-log vendor
- Tombstoning append-only no delete vendor
- Schema evolution free-form / no schema vendor
- Namespace primitives memory blocks per chat_store_key; user-scoped via chat_store + chat_store_key vendor
- Contradiction handling always append (no resolution) repo
- Forgetting policy consolidation/summarization vendor
- MCP support via official adapter — llama-index-tools-mcp vendor
- A2A support supported — LlamaIndex A2A agent server repo
- OTel via adapter — Arize Phoenix / OTel vendor
- Import / export JSON; multiple vector store backends vendor
- Integration count 0 repo
- Orchestration single-agent (Memory module per ChatEngine/AgentRunner) vendor
- Programmatic control full CRUD via Python/TS SDK vendor
- Pricing specifics OSS free; LlamaCloud tier flat-rate vendor
- Agent abstraction ReAct + state-machine (Workflows) vendor
- Memory primitives multiple (FIFO chat + vector long-term blocks) vendor
- LLM lock any (BYO) vendor
- Runtimes Python, Node/TypeScript vendor
- Session handling persistent thread (Memory class) vendor
- Validated verticals RAG developers, agent developers (horizontal) vendor
- Time to running <10min (pip install llama-index) vendor
- Anti-fit not for non-Python stacks vendor
- Optimised for FIFO + Memory Blocks for chat agent context vendor
- Adjacent infrastructure requires LlamaIndex; BYO LLM + vector store vendor
- Pros Mature retrieval ecosystem behind it — memory benefits from years of LlamaIndex retrieva… vendor
- Cons Retrieval-tilted — memory feels like a RAG add-on rather than a first-class abstraction. vendor
- Project & sources GitHub run-llama/llama_index vendor
- commit-trajectory [{"ym":"2025-07","cum":6890},{"ym":"2025-08","cum":7000},{"ym":"2025-09","cum":7118},{"y… repo
- download-trajectory [{"ym":"2025-11","monthly":3323391,"cum":3323391},{"ym":"2025-12","monthly":6965656,"cum… registry
- cost-input-usd-per-mtok 0 vendor
- cost-output-usd-per-mtok 0 vendor
- cost-tier free vendor
- cost-pricing-model free-tier vendor
- cost-last-verified 2026-06-29 vendor
- adoption-accelerator aws-genai-accelerator-2025 third-party
- architecture-deployment-locus local vendor
- licensing-bucket open-core vendor
- positioning-platform-vs-product platform vendor
- taxonomy-three-layer-rollup agent-infra repo
- developer-activity-opendigger {"activity":548.37,"as_of":"2026-09","openrank_12mo":28.66,"openrank_current":26.99,"par… third-party
- sdk-languages python,typescript repo
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
Related systems
References (1)
- LlamaIndex (framework) depends on at runtime — adjacent-infrastructure cell: requires LlamaIndex; BYO LLM + vector store