Khanmigo
Stores full chat transcripts per student (visible to learner and teacher via dashboard). Layers over Khan Academy's existing mastery graph (exercise completion, skill progress) for cross-session topic awareness; cross-session memory is derived from structured mastery state rather than free-text recall.
At a glance
- Type
- Chat history + Khan Academy mastery state
- Tier
- T2
- Created
- 2023 (Khan Academy; Khanmigo launched March 15 2023 with GPT-4)
- Latest release
- not applicable — not OSS
- License
- not applicable — not OSS
- GitHub
- not applicable — no GitHub repo
- Pricing
- Free + paid
- Funding
- Non-profit (Khan Academy); no funding rounds
Taxonomy
- storage
- relational
- retrieval
- injection
- persistence
- cross-session
- update
- append-only
- unit
- episode
- governance
- auditable
- conflict
- append
When to use
Optimised for: learner profile + spaced repetition + assessment-driven adaptation
Anti-fit: not for non-educational use cases
Pros & cons
Pros
Uniquely integrates AI tutoring with a structured curriculum graph — memory is mapped to a known knowledge domain, not just chat history.
Cons
Adoption is shallow — only ~15% of students with access actually use it; cross-session memory is bounded by mastery taxonomy rather than open-ended learning goals.
Claims & capabilities
2M students/educators/parents in 2024–25 school year. K–12 student usage 40k → 700k YoY (Khan Academy SY24-25 Annual Report).
Technical surface
- API surface
- searched not found
- Backend storage
- searched not found
- Deployment
- Managed-only
- Embedding model
- searched not found
- Multi-tenancy
- searched not found
- MCP
- no MCP support advertised — vertical product, no MCP server / client integration documented
- A2A
- no A2A protocol support advertised — vertical product, no A2A integration documented
- OpenTelemetry
- no OpenTelemetry integration advertised — vendor logs/observability not publicly documented
Similar systems
Other vertical / domain-specific ai memory in the catalog, ranked by inbound references.
- NVIDIA ReMEmbR T3
Builds long-horizon memory by captioning video segments with VILA, storing captions with timestamps + 3D position coordinates in MilvusDB. At query time, LLM iterates retrieval across text, time, and position modalities. Deployed on Nova Carter robot (Jetson Orin).
- Abridge T1
Clinician-assist ambient documentation. Source mapping: every AI-generated summary element traced back to the source utterance. Audit-and-trust layer over episodic memory. Built on proprietary 1.5M+ medical-encounter dataset.
- ASAPP GenerativeAgent T1
Treats memory as first-class architecture. Captures the digital footprint of every interaction; retrieves preference and history at engagement time. Public example: airline knowing a frequent flyer wants aisle seats with her son — preference-aware, not just history-lookup.
- BenevolentAI T1
Target identification / drug repurposing / mechanism tracing. 85+ data sources, petabyte-scale, rebuilt every few weeks. Wet-lab results re-enter the graph and shift downstream predictions — institutional experimental memory closing a feedback loop.
- Causaly T1
Drug discovery / target identification / causal mechanism tracing. The graph is the memory: 7 years of curated biomedical cause-effect relationships compounding with each new ingestion. Scientific RAG retrieves from a versioned causal substrate.
- Character.ai T1
Chat Memories (user-defined facts), auto-memories for c.ai+ subscribers, pinned memories, in-context retention. PipSqueak 2 model (April 2026) reduces in-conversation drift. Memory Visualization meter forthcoming.