Encrypted in transit and at rest.
Memory content is encrypted as it moves between your device and Anuma, and while it is stored. It is not end-to-end encrypted.
How Anuma
Private Memory Works
Memory Engine handles automatic recall. Memory Vault stores your curated knowledge. Together they give every AI model complete context.
Anuma Memory Engine
Automatic recall from conversations. Searches past conversations using semantic similarity. No exact keywords needed. Finds relevant context from days or weeks ago.
Learn moreAnuma Memory Vault
Persistent facts and preferences. A curated knowledge base of saved facts, preferences, and instructions. Explicitly saved. Available across all future conversations.
Learn moreHow Anuma AI memory works
Four steps from conversation to recall
Every memory follows the same path. Captured as you chat. Encrypted in transit and at rest. Backed up only if you want. Retrieved only when you need it.
- 01
Capture
As you chat, Anuma quietly captures meaningful moments: messages worth recalling and facts worth saving.
Every message is chunked at sentence boundaries (~400 chars, 50-char overlap).
- 02
Encrypt
Your memory is encrypted in transit and at rest, so it is protected as it moves between your device and Anuma.
Not end-to-end encrypted. Anuma and its service providers may access it for operation, safety and legal compliance.
- 03
Store
Encrypted backups sync your memory across your devices and are held on Anuma servers.
Turn backups off in Settings and the server copy is erased. You can delete your data any time.
- 04
Retrieve
When you ask anything in any AI model, Anuma finds the relevant memory and quietly threads it into context.
Semantic search finds meaning, not keywords. Your prompt goes to the model provider you choose.
Anuma Memory Engine
Automatic recall from every conversation you have ever had.
Semantic search.
Every message is embedded as a high-dimensional vector capturing its meaning. Queries find the most similar past messages using cosine similarity.
No exact keyword matches needed. A question about “deployment issues” surfaces past conversations about “CI pipeline failures.”
Vector embeddingsCosine similarity0.3 threshold
Query
“deployment issues”
Embed & compare vectors
Match · 0.87
“CI pipeline failures”
Smart chunking.
Long messages automatically split into overlapping segments using sentence-boundary detection. Each chunk targets roughly 400 characters with 50-character overlap.
Each chunk is embedded independently. Relevant fragments inside long messages still surface in search.
Sentence boundaries~400 char chunks50 char overlap
Input
Long message · 1,200 characters
Output
Chunk 1
Chunk 2
Chunk 3
~400 chars each · 50-char overlap
Diverse results.
Round-robin deduplication ensures results come from multiple conversations, not just one. The system over-fetches 3x to 9x candidates before deduplication for a diverse result pool.
Round-robin dedup3x to 9x over-fetchMulti-conversation
72
candidates fetched at 9x
Round-robin deduplication
8
results from 8 different conversations
Context expansion.
Finding a single chunk is not enough. The agent retrieves surrounding messages for full context. Default returns the full conversation session around each match.
Results are grouped by conversation with relevance scores and timestamps.
Full session contextRelevance scoresTimestamps
Previous message
Previous message
Matched chunk0.91
Following message
Following message
Anuma Memory Vault
A knowledge base that grows with you.
Save, update, and organize.
The agent creates, updates, and organizes memories into folders. Each memory is scoped for access control. Mark memories as private or shared.
Memories are filterable by folder or scope. Unfiled entries are queryable separately.
FoldersPrivate or sharedFilterable
Work preferences4 items
Personal7 items
Health goals3 items
Private to you
Semantic search with caching.
Embedding-based retrieval by meaning, not keywords. LRU cache holds up to 5,000 vectors. Evicts least-accessed entries when full.
New memories are embedded eagerly and cached immediately. The system gets faster over time.
Embedding retrievalLRU cache5,000 vectors
Vector capacity
68%
3,412 / 5,000 vectors
12ms
Avg retrieval
5,000
LRU cache limit
Supersession detection.
When two memories are semantically similar (70%+ cosine similarity) and 30+ days apart, the newer one replaces the older. Greedy pairing prevents cascading adjustments.
Original memories stay intact. Ranking is adjusted at query time only.
70% similarity30-day thresholdNon-destructive
Jan 2
“Prefers light mode”
82% similar · 45 days
Feb 16
“Prefers dark mode”
Supersedes older memory
User control.
Confirmation callbacks let users approve what the agent saves. Multi-user support with user-scoped storage enforced at the database layer.
Bulk delete available for full data cleanup.
Approval callbacksUser-scopedBulk delete
Agent wants to save
“User prefers dark mode in all editors”
Approve
Reject
User-scoped · you control what is saved
Better together.
Memory Engine and Memory Vault share the same embedding infrastructure. They serve different roles but work side by side.
1
User mentions they prefer dark mode in a conversation.
2
Memory Engine indexes this conversation automatically.
3
Agent uses Memory Vault to save “User prefers dark mode” as a persistent fact.
4
Future conversations: Vault retrieves the preference instantly without searching history.
| Memory Engine | Memory Vault | |
|---|---|---|
| Source | Conversation history | Explicitly saved memories |
| Updates | Automatic | Save/update via agent |
| Organization | By conversation | By folder and scope |
| Search default | 8 results, 0.3 threshold | 5 results, 0.1 threshold |
| Chunking | Sentence-based, ~400 chars | Whole memory as single unit |
| Staleness handling | Round-robin dedup | Supersession detection |
| Caching | Optional per-embedding | LRU cache (5,000 entries) |
| Best for | Recalling past discussions | Persistent facts and preferences |
Architecture
Encrypted in transit and at rest. Backups you control.
Your memory is stored on your device, with encrypted backups held on Anuma servers. It is encrypted in transit and at rest, but not end-to-end encrypted, so it may be accessible to Anuma and its service providers for operation, safety and legal compliance.
Encrypted in transitEncrypted at restBackups you controlCross-model
Stored on your device too.
Your memory also lives in a database on your device, so it is available wherever you use Anuma.
Backups you can turn off.
Encrypted backups of your conversations are held on Anuma servers. Turn backups off in Settings and the server copy is erased.
Unified across models and devices.
Your memory follows you across ChatGPT, Claude, Gemini, and every model on Anuma, and across every device you sign in from.
Data flow
Where your memory lives.
Your device
Memory stored in an on-device database
Anuma servers
Encrypted backup. Not end-to-end. Turn off to erase.
Your other devices
Sync from your encrypted backup
Memory is encrypted in transit and at rest. Anuma and its service providers may access it for operation, safety and legal compliance, and you can turn backups off at any time.
Traditional AI memory vs Anuma
Same job. Two approaches to AI memory.
Traditional AI memory lives on vendor servers, locked to one model. Anuma's AI memory is encrypted in transit and at rest, unified across every AI model, and yours to control.
Traditional AI memory
Cloud-onlyVendor-controlledModel-locked
Anuma memory
Device + backupUser-controlledPortable
Storage location
Traditional AI memory: Vendor cloud
Locked inside one provider, beside billions of other users.
Anuma memory: Device + backup
Stored on your device, with encrypted backups on Anuma servers you can turn off.
Encryption at rest
Traditional AI memory: Vendor-managed
Encrypted, with keys managed by the vendor.
Anuma memory: In transit and at rest
Encrypted, but not end-to-end encrypted.
Who can access it
Traditional AI memory: The vendor
The vendor and its service providers.
Anuma memory: Anuma and providers
May access data for operation, safety and legal compliance.
Cross-model portability
Traditional AI memory: Locked to one model
Memory works in one assistant. Switch tools and it is gone.
Anuma memory: Every model
One memory layer across ChatGPT, Claude, Gemini, and more.
User-controlled deletion
Traditional AI memory: Maybe, sometimes
Deletion requests, retention windows, and backup ambiguity.
Anuma memory: In the app
Delete memories or conversations, turn off backups to erase the server copy, or delete your account.
Used for training
Traditional AI memory: Often by default
Your conversations feed the next model release. Opt-out only.
Anuma memory: Not by default
Anuma does not use your memory to train its own models by default.
Audit transparency
Traditional AI memory: Opaque
Closed-source pipelines. You take the policy on faith.
Anuma memory: Inspectable
Open architecture. The crypto and data flow are documented.
AI persistent memory privacy
What Anuma keeps, and what you control.
Plain language about what Anuma holds, what goes to third parties, and the controls you have over each.
Encrypted conversation backups.
Held on Anuma servers, encrypted but not end-to-end encrypted. Turn backups off in Settings and the server copy is erased.
Messages and memory, encrypted.
Encrypted in transit and at rest. They may be accessible to Anuma and its service providers for operation, safety and legal compliance.
Prompts sent to model providers.
Prompts and inputs go to third-party providers. Their retention and training follow their own terms, which Anuma does not control.
Memory visibility, set by you.
Each memory can be private, matchable or public. Private memories are excluded from matching and public profiles.
No ad-targeting profiles.
We do not build advertising profiles from your conversations.
Deletion in your hands.
Delete memories or conversations, unpublish shared links, disconnect connected apps, or delete your account in the app.
The details are in writing. Our Privacy Policy describes how your data is collected, used and protected, and the controls above are available in the app.
Memory questions.
Everything you need to know about how Anuma remembers.
Ask us anythingMemory Engine automatically indexes and searches past conversations. Memory Vault stores specific facts and preferences that are explicitly saved.
Memory Engine indexes everything automatically. For Memory Vault, the AI suggests what to save, but you can approve or reject each memory.
Every message and memory is converted into a vector that captures its meaning. Search finds the most similar vectors using cosine similarity, so you do not need exact keyword matches.
Yes. Memory Vault supports folders and scopes. You can filter searches by folder, mark memories as private or shared, and move entries between folders.
Memory Vault uses supersession detection. When a newer memory is semantically similar to an older one and enough time has passed, the newer version is automatically prioritized in search results.
Yes. Memory is encrypted in transit and at rest. It is not end-to-end encrypted, so it may be accessible to Anuma and its service providers for operation, safety and legal compliance.
Yes. You can delete individual memories or do a full bulk delete from your account, and you can delete your account in the app from Settings > Account > Delete account.
Yes. Your memory loads into every conversation regardless of which model you use. Memory Engine and Memory Vault work with ChatGPT, Claude, Gemini, and all other models on Anuma.
There is no hard limit. The Memory Vault cache holds 5,000 vectors for fast access. Memories beyond the cache are still searchable but may take slightly longer on first access.
Give your AI a memory that belongs to you.
Memory Engine and Memory Vault give every AI model complete context, with controls to delete it whenever you want.