Roadmap

Our vision for AI memory and what's coming next. Have ideas? Open an issue.

Completed
In Progress
Planned
Researching
v1

Core Memory

Completed

remember tool

Store memories with vector embeddings and semantic deduplication

recall tool

Semantic search with tag filters

forget tool

Archive to KV before deletion

list tool

Paginated memory listing

Auth & API keys

SHA-256 hashed keys with namespace isolation

Daily decay

Automatic strength reduction and archival

Weekly consolidation

Cluster similar memories with Claude (Pro)

v2

Memory Graph

Completed

link tool

Typed relationships between memories

Auto-linking

Automatic connection of similar memories

Graph expansion

Include neighbors in recall results

Hybrid search (RRF)

Combine semantic + graph with Reciprocal Rank Fusion

update tool

Modify existing memories

Full-text search

tsvector column for keyword search

v2.1

Compliance & Security

Completed

audit_log

Immutable operation records (GDPR Art. 30)

tos_acceptances

Versioned consent tracking (GDPR Art. 7)

gdpr_requests

Data subject request management

stats tool

Namespace usage counters

Rate limiting

Monthly limits per plan

RLS policies

Row-level security on all tables

v2.2

CoT Feedback

Completed

cot_feedback tool

Record Chain of Thought effectiveness

CoT-boosted recall

Prioritize memories with proven effectiveness

CoT decay

Stable decay for feedback scores (0.02/day)

Session tracking

Link memories to CoT sessions

v2.3

Enhanced Dashboard

In Progress

Real-time stats

Live memory and usage metrics

Memory explorer

Visual graph of memory relationships

Search interface

Test recall directly from dashboard

API key management

Create, revoke, and monitor keys

Team collaboration

Share namespaces with team members

v3

Platform & Scale

Planned

LemonSqueezy integration

Pro plan billing and subscriptions

Memory import/export

Bulk operations and backups

Webhooks

Real-time events for memory changes

Memory analytics

Insights into usage patterns

Custom embeddings

Support for other embedding models

v4

Advanced Memory

Planned

Multi-modal memory

Images, audio, and video support

Hierarchical memory

Projects, folders, and nested organization

Memory templates

Structured memory types (contacts, tasks, etc.)

Temporal queries

"What did I know in January?"

Cross-namespace search

Search across multiple namespaces

Future

Research

Planned

Federated memory

Share memories between trusted agents

Differential privacy

Privacy-preserving memory aggregation

Local-first sync

Edge caching with conflict resolution

Memory compression

LLM-based memory summarization

Shape the future

Have a feature request or idea? We'd love to hear it.

🚀 Launch in seconds

Ready to give your AI
a persistent memory?

Join thousands of developers building smarter agents with takizen.

10k+ Agents connected
2M+ Memories stored
<10ms Avg. latency