Memory

Preview — See Feature availability. Memory APIs and workflow hooks are implemented. The dedicated Memory nav in the Console is off in production (enableMemory false). Use the Memory API, Platform MCP, and Autopilot memory tools in enabled environments.

Memory is AgentRuntime’s long-term knowledge layer. It extracts facts from conversations, files, and workflow step outputs, then makes them searchable for agents and workflows. Chat remembers the current thread. Memory remembers what matters across sessions.

What memory stores

Source	Episode type	How it gets in
Chat conversations	`chat_conversation`	Index conversation API or auto-index on save
Work files	`work_file`	Index file API
Workflow step output	`run_step`	Runtime memory hooks on step completion
Manual / scripts	`external_event`	Direct extract API

Extracted knowledge is scoped by tenant and project.

When to use memory

Use case	Approach
Agent recalls past decisions	Search memory before LLM steps (prepare-step or explicit search)
Onboard context from chat history	Index key conversations after milestones
Document Q&A across uploads	Index uploaded files, search at query time
Workflow learns from prior runs	`run_step` episodes capture step outputs automatically

Memory complements — does not replace — workflow state. Step results in {{steps.*}} templates are run-scoped. Memory persists across runs.

Core operations

Search

Semantic search over confirmed long-term memory:

POST /v1/memory/search

{
  "query": "What did we decide about Q2 pricing?",
  "limit": 10,
  "project_id": "your-project-id"
}

Also available as Platform MCP tool memory_search. Requires project_contributor.

Index a conversation

Pull a chat transcript and queue extraction:

POST /v1/memory/index-conversation

{
  "conversation_id": "conv-uuid",
  "project_id": "your-project-id"
}

Returns a job ID. Poll GET /v1/memory/extract/{jobId} for status.

Index a file

Index an uploaded work file for extraction:

POST /v1/memory/index-file

{
  "file_id": "file-uuid",
  "project_id": "your-project-id"
}

Batch extraction

For large backfills, use batch mode:

POST /v1/memory/extract/batch
GET /v1/memory/extract/batch/{batchId}

Bulk jobs run on a separate queue and may take longer than interactive indexing.

Get an episode

Retrieve a stored episode bundle:

GET /v1/memory/episodes/{episodeId}

Add ?detail=false to omit item payloads.

Agent memory tools

Agents and Autopilot can call memory tools via:

POST /v1/memory/tools/call

Tool	Purpose
`memory_search`	Semantic search
`memory_get_episode`	Fetch episode by ID
`memory_get_continuation`	Latest open episode for a conversation
`memory_propose_items`	Propose candidate memory items
`memory_confirm_item`	Promote item to confirmed long-term memory

Platform MCP exposes the same tools when the memory group is enabled.

Workflow integration

The workflow runtime can call prepare-step memory hooks before LLM steps, injecting retrieved context into the prompt. If memory is unavailable, the run continues without LTM context (logged as a warning, not a hard failure). Pattern:

Index relevant conversations/files into memory
LLM workflow steps benefit from prepare-step context automatically when configured
Optionally add an explicit memory_search via Platform MCP or a future builtin tool step

Billing

Memory actions consume credits:

Action	Typical cost
`memory_extract_interactive`	~10 credits per job
`memory_extract_bulk`	~2 credits per job
`memory_index_file_mb`	~1 credit per MiB

See Billing and credits.

Auto-indexing (when enabled)

When memory is enabled for your workspace, AgentRuntime can automatically index chat conversations as they happen, so later searches and workflow steps can draw on that context without an explicit indexing call.

Auto-indexing behavior is configured per workspace by AgentRuntime — it is not yet a Console toggle. Contact support if you want auto-indexing enabled or adjusted for your workspace.

Limitations

No dedicated Memory browser in production Console yet
Search quality depends on indexed content — empty memory returns empty results
Extraction is async — poll job status before assuming facts are available
Memory is tenant/project isolated — no cross-tenant search

Troubleshooting

Issue	Fix
`503 memory kernel not configured`	Memory service not deployed in this environment
`queue unavailable`	Extract worker or Redis queue down — contact support
`work file not found`	Verify `file_id` exists and the file is still available
Search returns nothing	Confirm indexing job completed (`GET /v1/memory/extract/{jobId}`)
High credit usage	Prefer bulk extraction for large backfills; set indexing scope narrowly

See Troubleshooting.

Autopilot and chat — chat threads that feed memory
Platform MCP — memory_search and related tools
Workflow patterns — memory-enriched LLM steps

Get started

Platform

Workflows

Guides

Integration setup

Connectors

AI & chat

API

What memory stores

When to use memory

Core operations

Search

Index a conversation

Index a file

Batch extraction

Get an episode

Agent memory tools

Workflow integration

Billing

Auto-indexing (when enabled)

Limitations

Troubleshooting

​What memory stores

​When to use memory

​Core operations

​Search

​Index a conversation

​Index a file

​Batch extraction

​Get an episode

​Agent memory tools

​Workflow integration

​Billing

​Auto-indexing (when enabled)

​Limitations

​Troubleshooting

​Related docs

What memory stores

When to use memory

Core operations

Search

Index a conversation

Index a file

Batch extraction

Get an episode

Agent memory tools

Workflow integration

Billing

Auto-indexing (when enabled)

Limitations

Troubleshooting

Related docs