Spellbook Library
Transactional / contract drafting. Ingests entire precedent corpus (OneDrive, Dropbox); ML surfaces and reuses the firm's own prior clauses during live drafting in Word, tuned to specific deal context. Goal: eliminate "reads like ChatGPT" output.
At a glance
- Type
- Firm-precedent corpus memory
- Tier
- T1
- Created
- 2018 (company founded 2018; Spellbook product launched 2022; Toronto Canada)
- Latest release
- not applicable — not OSS
- License
- not applicable — not OSS
- GitHub
- not applicable — no GitHub repo
- Pricing
- searched not found
- Funding
- $31M total Series A · 2024-01
Taxonomy
- storage
- vector
- retrieval
- similarity
- persistence
- long-term
- update
- extraction
- unit
- document
- governance
- auditable
- conflict
- none
When to use
Optimised for: matter-scoped privilege + audit + firm-precedent memory
Anti-fit: not for non-legal verticals; not for self-represented litigants
Pros & cons
Pros
Lowest-friction legal AI — sits inside Word with no separate UI; firm-knowledge library is curated, not opaque.
Cons
Word-only; transactional drafting focus; less depth than Harvey on litigation or research workflows.
Claims & capabilities
$50M Series B (Khosla, Oct 2025). $350M post-money. ~$80M total funding.
Technical surface
- API surface
- searched not found
- Backend storage
- searched not found
- Deployment
- searched not found
- Embedding model
- searched not found
- Multi-tenancy
- Multi-tenant SaaS with zero data retention from LLM providers (data exists only in memory for request)
- 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.