Harvey Memory
https://www.harvey.ai/blog/memory-in-harvey
Memory carries forward user-chosen context (matter details, precedent, preferences, approved practices) and reuses past threads scoped to a user, client, matter or firm process. Firm admins enable/disable it; ethical walls and audit controls promised. Announced Jan 2026 as co-development with customers.
Active · Verified 2026-10-22 · fresh as of 2026-10-22 · 61 cited facts, 6 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Type
- 4-tier matter / firm / client-institution memory vendor fix
- Tier
- T1
- Latest release
- not applicable — not OSS fix
- License
- not applicable — not OSS fix
- GitHub
- not applicable — no GitHub repo fix
- Pricing
- Enterprise only third-party fix
- Funding
- latest: $550M at $15.5B valuation (co-led by Lightspeed and Diffusion, 2026-09); prior $200M at $11B (2026-03) vendor fix
- Freshness
- 2026-10-22 fix
- Layer
- agent-infra repo fix
- Licensing
- commercial third-party fix
Memory taxonomy
- storage
- vector
- retrieval
- similarity
- persistence
- long-term
- update
- extraction
- unit
- episode
- 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
Most established legal AI brand — Harvey Memory adds firm + matter + drafted-document recall on top of mature legal model.
Cons
Closed enterprise product; pricing tier excludes solo / small-firm market; firm-curated corpus required to get value.
Claims & capabilities
$190M ARR, 1,000+ customers at announcement.
Technical surface
- API surface
- searched not found
- Backend storage
- searched not found
- Deployment
- Both
- Embedding model
- searched not found
- Multi-tenancy
- Logical workspace separation: every user mapped to unique workspace; workspaces cannot communicate with each other (no cross-workspace access)
- MCP
- yes — Harvey acts as MCP client via MCP Connector Library (vendors/firms plug tools in), with an MCP policy engine; own MCP server only planned (Dec 2025)
- A2A
- no A2A protocol support advertised — vertical product, no A2A integration documented
- OpenTelemetry
- no OpenTelemetry integration advertised — vendor logs/observability not publicly documented
Alternatives to Harvey Memory → live options ranked by shared design, licence, hosting and MCP support
Similar systems
Other vertical / domain-specific ai memory in the catalog, ranked by inbound references.
- Inworld AI T1
Formerly an NPC character engine with named Long-Term Memory; now sells realtime voice models (TTS-2, STT-1, speech-to-speech) and OpenAI-compatible open-model inference. Character/NPC memory is no longer the headline product.
- 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
Taken private: merged into Osaka Holdings (renamed BenevolentAI S.A.) and delisted from Euronext Amsterdam 2025-03-13. 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. As of Oct 2026 benevolent.com leads with Qor, powered by BenevolentAI, a life-science intelligence service for biopharma decision makers and consultancies. Qor (qor.one) is a continuously updated decision-intelligence service built on BenevolentAI's knowledge graph and reasoning methods: it tracks evidence on an asset or indication (competitive position, rival readouts, comparable deals), source-verifies each claim and flags when an earlier conclusion stops holding. It targets biotech teams and life-science consultancies.
- Causaly T1
Agentic AI platform for life sciences: orchestrated scientific agents over a curated biomedical knowledge graph (plus Bio Graph and Pipeline Graph), an evidence-retrieval system and an enterprise data fabric over literature, trials and patents.
Sources (60 cited facts, 6 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is 4-tier matter / firm / client-institution memory vendor
- What it does & what's distinct Memory carries forward user-chosen context (matter details, precedent, preferences, appr… vendor
- Claims / benefits $190M ARR, 1,000+ customers at announcement. vendor
- Modalities text vendor
- Created 2024-07 third-party
- Funding latest: $550M at $15.5B valuation (co-led by Lightspeed and Diffusion, 2026-09); prior $… vendor
- Customers / scale $195M ARR 201-500 ppl A&O Shearman, Macfarlanes, Mori Hamada & Matsumoto, A&L Goodbody vendor
- Pricing Enterprise only third-party
- Compliance SOC 2 Type II, ISO 27001, GDPR-compliant, EU-US DPF vendor
- Data handling Never trains on customer data (contractual commitment to all AI providers: no retention,… vendor
- Deployment Both vendor
- HQ US third-party
- Memory volume / scale internal — undisclosed (1000+ enterprise clients) vendor
- Multi-tenancy Logical workspace separation: every user mapped to unique workspace; workspaces cannot c… vendor
- Encryption At-rest + in-transit (strong access controls, no training on customer data by default); … vendor
- Consistency strong consistency (legal accuracy requirement) vendor
- Versioning matter-versioned (per-matter document history) vendor
- Tombstoning matter-closure soft delete + privilege-log retention vendor
- Schema evolution schema with migration vendor
- Namespace primitives matter-per-tenant + firm-tenant (ethical-wall isolation) vendor
- Contradiction handling lawyer-in-the-loop reconciliation; auditor flags vendor
- Forgetting policy matter-retention policy (firm-defined; commonly 7-10 years post-matter-close per state b… vendor
- MCP support yes — Harvey acts as MCP client via MCP Connector Library (vendors/firms plug tools in),… vendor
- Import / export proprietary only (enterprise) vendor
- Validated verticals legal (BigLaw, plaintiff litigation, in-house, transactional, contracts) vendor
- Time to running <1week to weeks-to-months (firm procurement + matter integration) vendor
- Anti-fit not for non-legal verticals; not for self-represented litigants vendor
- Optimised for matter-scoped privilege + audit + firm-precedent memory vendor
- Adjacent infrastructure requires DMS / case-management integration + identity provider vendor
- Pros Most established legal AI brand — Harvey Memory adds firm + matter + drafted-document re… vendor
- Cons Closed enterprise product; pricing tier excludes solo / small-firm market; firm-curated … vendor
- Project & sources memory blog vendor
- obs-langsmith no vendor
- obs-opentelemetry no vendor
- obs-datadog no vendor
- obs-helicone no vendor
- obs-weave no vendor
- obs-langfuse no vendor
- obs-arize no vendor
- obs-custom Closed enterprise legal-AI product; no published observability surface vendor
- cost-token-budget no vendor
- cost-prompt-caching no vendor
- cost-semantic-caching no vendor
- cost-batching no vendor
- cost-model-routing no vendor
- cost-streaming-only no vendor
- cost-observability-cost-attribution no vendor
- eval-langsmith-evals no vendor
- eval-braintrust no vendor
- eval-weights-and-biases-agent no vendor
- eval-helicone-evals no vendor
- eval-custom-test-harness yes vendor
- eval-human-loop yes vendor
- eval-production-traffic-replay no vendor
- cost-pricing-model subscription third-party
- cost-last-verified 2026-06-29 third-party
- architecture-deployment-locus cloud vendor
- licensing-bucket commercial third-party
- positioning-platform-vs-product product vendor
- taxonomy-three-layer-rollup agent-infra repo
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
- 2026-10-22
- Funding / ownership “$1.1B total $11.0B val Growth · 2026-03” → “latest: $550M at $15.5B valuation (co-led by Lightspeed and Diffusion, 2026-09); prior $200M at $11B (2026-03)” source
- MCP support first recorded as “yes — Harvey acts as MCP client via MCP Connector Library (vendors/firms plug tools in), with an MCP policy engine; own MCP server only planned (Dec 2025)” source