Agents that remember what matters.

Give your agents persistent, importance scored memory across every session. Facts, preferences, and constraints learned in one run are automatically recalled in the next.

Get started freeTalk to sales

Learn

During each run, agents write important facts, preferences, and constraints into their memory workspace — automatically, from tool calls and LLM responses.

tool_callUser prefers named exports and strict mode0.95
tool_callRate limit: 100 req/min on external API0.67
llm_responseSummarise reports as markdown tables0.52

Recall

At the start of each subsequent run, the most relevant memories are surfaced automatically — ranked by importance and recency — so agents always have context.

Recalling memories for code_agent…
0.95TypeScript strict mode + named exports14×
0.81Production DB: reads → replica, writes → primary8×
0.38HTTP timeout: 30s across all services3×

Manage

Browse, search, pin, edit, and delete memories from the dashboard or CLI. Memories are scoped to workspaces and organised by project.

Search memories…
myproject/main247
myproject/staging83
analytics/prod61

Context that compounds.

The Memory Plugin gives any agent a persistent workspace. Memories are scored by importance, critical constraints surface first, while low-priority notes stay available when needed. Pinned memories are always recalled.

✓Importance scoring from 0.0–1.0 with colour-coded badges
✓Source tracking: know whether a memory came from a tool call or LLM response
✓Pin high value memories to guarantee recall on every run
✓Per-agent attribution across shared workspaces
✓Edit or delete individual memories from dashboard or CLI
MemoriesDaily Logs
myproject/main (247) ▾

Total Memories

247

Across 3 workspaces

Storage Used

1.8 MB

Across 3 workspaces

Last Updated

Today

Sep 1 → Today

ContentSourceAgentPinImportance

User prefers TypeScript strict mode and named exports across all modules

tool_callcode_agent
0.95

Production DB uses read-replicas; write operations must target primary

llm_responsedata_agent
0.81

Rate limit on external API is 100 req/min — add exponential backoff

tool_callapi_agent
0.67

Summarise all reports as markdown tables with one-line executive summary

llm_responsereport_agent
0.52

Default timeout for all HTTP requests set to 30s across services

tool_callinfra_agent
0.38

A living record of everything your agents have learned.

Alongside discrete memories, agents write daily logs that capture the full narrative of what happened in each session. A GitHub-style activity heatmap shows memory growth over time — so you can see exactly how your agents are evolving.

✓Daily log entries written automatically after each agent run
✓Searchable log history with full session context
✓Recall count tracking: see which memories are used most
✓Workspace-scoped storage with per-project isolation
Daily Logs

Activity Calendar

84 daily logs

Less
More
Sep
Oct
Nov
Dec
Jan
Feb
Mon
Wed
Fri
84
Total logs
61 / 126
Active days
9 logs
Peak day

Manage memory from your terminal

List, search, pin, and inspect memories without leaving the CLI.

daita cli
# List memories for a workspace
$ daita memory list --workspace myproject/main
  247 memories  •  1.8 MB  •  last updated today
 
# Search memories by content
$ daita memory search "TypeScript" --workspace myproject/main
  [0.95] User prefers TypeScript strict mode and named exports
  [0.74] All interfaces should be defined in /src/types/**
 
# View memory stats across all workspaces
$ daita memory stats
  Total: 247 memories  •  3 workspaces  •  Storage: 1.8 MB
  Top agents: code_agent (91)  data_agent (74)  api_agent (52)
 
# Pin a high-value memory
$ daita memory pin mem_a3f9b2 --workspace myproject/main
  ✓ Memory pinned — will be included in all future recalls
daita memory listdaita memory searchdaita memory statsdaita memory pindaita memory delete

Give your agents a memory.

Add the Memory Plugin to any agent and start building context that compounds over time.

Get started freeTalk to sales