openclaw/docs/reference/memory-config.md

summary, title, sidebarTitle, read_when

summary	title	sidebarTitle	read_when
Built-in memory search providers, retrieval modes, and multimodal indexing	Memory configuration reference	Memory config	You want to configure memory search providers or embedding models You want to understand hybrid search, MMR, or temporal-decay defaults You want to enable multimodal memory indexing

This page lists every configuration knob for OpenClaw memory search. For conceptual overviews, see:

How memory works. Default SQLite backend. Search pipeline and tuning. Memory sub-agent for interactive sessions.

All shared memory settings live under top-level memory in openclaw.json. Search defaults use memory.search; per-agent search overrides use agents.entries.*.memory.search.

For the recommended personal-agent workflow, use `memory.search.rememberAcrossConversations`. Advanced Active Memory targeting, model, prompt, and latency controls live under `plugins.entries.active-memory`.

See Active Memory for both activation paths, transcript persistence, and safe rollout guidance.

---

Remember across conversations

Key	Type	Default	Description
rememberAcrossConversations	boolean	On for personal installs; off with configured DM isolation	Use relevant context from this agent's other recognized private conversations.

Configure it per agent when only a trusted personal agent should use cross-conversation transcript recall:

{
  agents: {
    entries: {
      personal: {
        memory: {
          search: {
            rememberAcrossConversations: true,
          },
        },
      },
    },
  },
}

The value follows normal memory.search inheritance with a per-agent override. When unset, it defaults on only if global session.dmScope is unset or "main" and no binding has a session.dmScope override. Any configured DM isolation defaults it off. An explicit true or false always wins. Enabling it implies session transcript indexing and adds sessions to the agent's resolved memory sources.

OpenClaw's built-in memory provider supports this protected path. Alternate memory providers can keep using their own recall hooks and advanced Active Memory tools, but this setting is skipped unless the current provider supports protected private transcript recall. openclaw doctor reports an unsupported provider or an explicit Active Memory toolsAllow list that omits memory_search.

The retrieval boundary is narrower than general session search:

• only the same agent's recognized private conversations are eligible
• the conversation being answered is excluded
• groups and channels are excluded as sources and destinations
• unknown conversation kinds fail closed
• sandboxed recall cannot use the special cross-conversation authorization

The setting does not change tools.sessions.visibility, session keys, transcript storage, delivery routing, or the permissions of sessions_list, sessions_history, and sessions_send. Active Memory performs a bounded read-only retrieval pass; unavailable or timed-out retrieval does not block the reply.

---

Provider selection

Key	Type	Default	Description
enabled	boolean	true	Enable or disable memory search
provider	string	"openai"	Embedding adapter ID such as bedrock, deepinfra, gemini, github-copilot, local, mistral, ollama, openai, openai-compatible, or voyage; may also be a configured models.providers.<id> whose api points at a memory embedding adapter or OpenAI-compatible model API
model	string	provider default	Embedding model name
fallback	string	"none"	Fallback adapter ID when the primary fails

When provider is not set, OpenClaw uses OpenAI embeddings. Set provider explicitly to use Bedrock, DeepInfra, Gemini, GitHub Copilot, Mistral, Ollama, Voyage, a local GGUF model, or an OpenAI-compatible /v1/embeddings endpoint. Legacy configs that still say provider: "auto" resolve to openai.

Changing the embedding provider, model, provider settings, sources, scope, chunking, or tokenizer can make the existing SQLite vector index incompatible. OpenClaw pauses vector search and reports an index identity warning instead of automatically re-embedding everything. Rebuild when you are ready with `openclaw memory status --index --agent ` or `openclaw memory index --force --agent `.

When provider is unset, legacy provider: "auto" is present, or provider: "none" intentionally selects FTS-only mode, memory recall can still use lexical FTS ranking when embeddings are unavailable.

Explicit non-local providers fail closed. If you set memory.search.provider to a concrete remote-backed provider such as Bedrock, DeepInfra, Gemini, GitHub Copilot, LM Studio, Mistral, Ollama, OpenAI, Voyage, or an OpenAI-compatible custom provider, and that provider is unavailable at runtime, memory_search returns an unavailable result instead of silently using FTS-only recall. Fix the provider/auth configuration, switch to a reachable provider, or set provider: "none" if you want deliberate FTS-only recall.

Custom provider ids

memory.search.provider can point at a custom models.providers.<id> entry for memory-specific provider adapters such as ollama, or for OpenAI-compatible model APIs such as openai-responses / openai-completions. OpenClaw resolves that provider's api owner for the embedding adapter while preserving the custom provider id for endpoint, auth, and model-prefix handling. This lets multi-GPU or multi-host setups dedicate memory embeddings to a specific local endpoint:

{
  models: {
    providers: {
      "ollama-5080": {
        api: "ollama",
        baseUrl: "http://gpu-box.local:11435",
        apiKey: "ollama-local",
        models: [{ id: "qwen3-embedding:0.6b", name: "Qwen3 Embedding 0.6B" }],
      },
    },
  },
  memory: {
    search: {
      provider: "ollama-5080",
      model: "qwen3-embedding:0.6b",
    },
  },
}

API key resolution

Remote embeddings require an API key. Bedrock uses the AWS SDK default credential chain instead (instance roles, SSO, access keys, or a Bedrock API key).

Provider	Env var	Config key
Bedrock	AWS credential chain, or AWS_BEARER_TOKEN_BEDROCK	No API key needed
DeepInfra	DEEPINFRA_API_KEY	models.providers.deepinfra.apiKey
Gemini	GEMINI_API_KEY	models.providers.google.apiKey
GitHub Copilot	COPILOT_GITHUB_TOKEN, GH_TOKEN, GITHUB_TOKEN	Auth profile via device login
Mistral	MISTRAL_API_KEY	models.providers.mistral.apiKey
Ollama	OLLAMA_API_KEY (placeholder)	--
OpenAI	OPENAI_API_KEY	models.providers.openai.apiKey
Voyage	VOYAGE_API_KEY	models.providers.voyage.apiKey

Codex OAuth covers chat/completions only and does not satisfy embedding requests.

---

Remote endpoint config

Use provider: "openai-compatible" for a generic OpenAI-compatible /v1/embeddings server that should not inherit global OpenAI chat credentials.

Custom API base URL. Override API key. Extra HTTP headers (merged with provider defaults).

{
  memory: {
    search: {
      provider: "openai-compatible",
      model: "text-embedding-3-small",
      remote: {
        baseUrl: "https://api.example.com/v1/",
        apiKey: "YOUR_KEY",
      },
    },
  },
}

---

Provider-specific config

| Key | Type | Default | Description | | ---------------------- | -------- | ---------------------- | ------------------------------------------- | | `model` | `string` | `gemini-embedding-001` | Also supports `gemini-embedding-2-preview` | | `outputDimensionality` | `number` | `3072` | For Embedding 2: 768, 1536, or 3072 |

<Warning>
Changing model or `outputDimensionality` changes the index identity. OpenClaw
pauses vector search until you explicitly rebuild the memory index.
</Warning>

OpenAI-compatible embedding endpoints can opt into provider-specific `input_type` request fields. This is useful for asymmetric embedding models that require different labels for query and document embeddings.

| Key                 | Type     | Default | Description                                             |
| ------------------- | -------- | ------- | -------------------------------------------------------- |
| `inputType`         | `string` | unset   | Shared `input_type` for query and document embeddings   |
| `queryInputType`    | `string` | unset   | Query-time `input_type`; overrides `inputType`          |
| `documentInputType` | `string` | unset   | Index/document `input_type`; overrides `inputType`      |

```json5
{
  memory: {
    search: {
      provider: "openai-compatible",
      remote: {
        baseUrl: "https://embeddings.example/v1",
        apiKey: "${EMBEDDINGS_API_KEY}",
      },
      model: "asymmetric-embedder",
      queryInputType: "query",
      documentInputType: "passage",
    },
  },
}
```

Changing these values affects embedding cache identity for provider batch indexing and should be followed by a memory reindex when the upstream model treats the labels differently.

### Bedrock embedding config

Bedrock uses the AWS SDK default credential chain plus an OpenClaw-checked bearer token, so no API keys are stored in config. If OpenClaw runs on EC2 with a Bedrock-enabled instance role, just set the provider and model:

```json5
{
  memory: {
    search: {
      provider: "bedrock",
      model: "amazon.titan-embed-text-v2:0",
    },
  },
}
```

| Key                    | Type     | Default                        | Description                     |
| ---------------------- | -------- | ------------------------------- | -------------------------------- |
| `model`                | `string` | `amazon.titan-embed-text-v2:0` | Any Bedrock embedding model ID  |
| `outputDimensionality` | `number` | model default                  | For Titan V2: 256, 512, or 1024 |

**Supported models** (with family detection and dimension defaults):

| Model ID                                   | Provider   | Default Dims | Configurable Dims          |
| ------------------------------------------- | ---------- | ------------- | -------------------------- |
| `amazon.titan-embed-text-v2:0`             | Amazon     | 1024         | 256, 512, 1024             |
| `amazon.titan-embed-text-v1`               | Amazon     | 1536         | --                          |
| `amazon.titan-embed-g1-text-02`            | Amazon     | 1536         | --                          |
| `amazon.titan-embed-image-v1`              | Amazon     | 1024         | --                          |
| `amazon.nova-2-multimodal-embeddings-v1:0` | Amazon     | 1024         | 256, 384, 1024, 3072       |
| `cohere.embed-english-v3`                  | Cohere     | 1024         | --                          |
| `cohere.embed-multilingual-v3`             | Cohere     | 1024         | --                          |
| `cohere.embed-v4:0`                        | Cohere     | 1536         | 256, 384, 512, 768, 1024, 1536 |
| `twelvelabs.marengo-embed-3-0-v1:0`        | TwelveLabs | 512          | --                          |
| `twelvelabs.marengo-embed-2-7-v1:0`        | TwelveLabs | 1024         | --                          |

Throughput-suffixed variants (e.g., `amazon.titan-embed-text-v1:2:8k`) and region-prefixed inference profile IDs (e.g., `us.amazon.titan-embed-text-v2:0`) inherit the base model's configuration.

**Region:** resolved in this order: the `memory.search.remote.baseUrl` override, the `models.providers.amazon-bedrock.baseUrl` config, `AWS_REGION`, `AWS_DEFAULT_REGION`, then a default of `us-east-1`.

**Authentication:** OpenClaw checks for `AWS_ACCESS_KEY_ID` + `AWS_SECRET_ACCESS_KEY` or `AWS_BEARER_TOKEN_BEDROCK` first, then falls through to the standard AWS SDK default credential provider chain:

1. Environment variables (`AWS_ACCESS_KEY_ID` + `AWS_SECRET_ACCESS_KEY`), unless `AWS_PROFILE` is also set
2. SSO (only when SSO fields are configured)
3. Shared credentials and config files (`fromIni`, includes `AWS_PROFILE`)
4. Credential process (`credential_process` in the AWS config file)
5. Web identity token credentials
6. ECS or EC2 instance metadata credentials

**IAM permissions:** the IAM role or user needs:

```json
{
  "Effect": "Allow",
  "Action": "bedrock:InvokeModel",
  "Resource": "*"
}
```

For least-privilege, scope `InvokeModel` to the specific model:

```text
arn:aws:bedrock:*::foundation-model/amazon.titan-embed-text-v2:0
```

| Key | Type | Default | Description | | ----------------- | -------- | --------------- | ----------------------- | | `local.modelPath` | `string` | auto-downloaded | Path to GGUF model file |

Install the official llama.cpp provider first: `openclaw plugins install @openclaw/llama-cpp-provider`.
Default model: `embeddinggemma-300m-qat-Q8_0.gguf` (~0.6 GB, auto-downloaded). Source checkouts still require native build approval: `pnpm approve-builds` then `pnpm rebuild node-llama-cpp`.

Use the standalone CLI to verify the same provider path the Gateway uses:

```bash
openclaw memory status --deep --agent main
openclaw memory index --force --agent main
```

Cache placement and embedding context sizing are provider-owned. `openclaw memory status --deep` reports last-known llama.cpp backend, device, offload, requested-context, and timestamped memory facts after the runtime has loaded; passive status does not load a model.

Set `provider: "local"` explicitly for local GGUF embeddings. `hf:` and HTTP(S) model references are supported for explicit local configs (via node-llama-cpp's model resolution), but they do not change the default provider.

Indexing behavior

Memory engines own synchronization, batching, watch, and post-compaction indexing heuristics. OpenClaw keeps these behaviors enabled with maintained defaults rather than exposing per-install timing switches.

Hybrid search config

All under memory.search.query:

Key	Type	Default	Description
maxResults	number	6	Max memory hits returned before injection
minScore	number	0.35	Minimum relevance score to include a hit

Hybrid retrieval remains enabled. The builtin engine always applies a fixed 30-day recency half-life to dated daily notes and a fixed importance multiplier after hybrid relevance, then applies MMR diversity ordering with a fixed lambda of 0.7. MEMORY.md, USER.md, and other evergreen memory files do not decay. Nullable importance is neutral, so no migration or new tuning key is required for existing indexes.

Strong trigger matches on promoted, trusted entries can inject up to three compact memories on eligible interactive turns. Today, root MEMORY.md and USER.md are the curated eligible tier. Daily notes and transcripts are never auto-injected.

Full example

{
  memory: {
    search: {
      query: {
        maxResults: 6,
        minScore: 0.35,
      },
    },
  },
}

---

Additional memory paths

Key	Type	Description
extraPaths	Array<string | { path: string; pattern?: string }>	Additional directories or files to index

{
  memory: {
    search: {
      extraPaths: ["../team-docs", { path: "/srv/shared-notes", pattern: "runbooks/**/*.md" }],
    },
  },
}

Paths can be absolute or workspace-relative. Directories are scanned recursively for supported files. Object entries narrow a directory with a root-relative glob using / separators; direct file entries are indexed exactly. The builtin engine skips symlinks.

---

Multimodal memory (Gemini)

Index images and audio alongside Markdown using Gemini Embedding 2:

Key	Type	Default	Description
multimodal.enabled	boolean	false	Enable multimodal indexing
multimodal.modalities	string[]	--	["image"], ["audio"], or ["all"]
multimodal.maxFileBytes	number	10485760	Max file size for indexing (10 MiB)

Only applies to files in `extraPaths`. Default memory roots stay Markdown-only. Requires `gemini-embedding-2-preview`. `fallback` must be `"none"`.

Supported formats: .jpg, .jpeg, .png, .webp, .gif, .heic, .heif (images); .mp3, .wav, .ogg, .opus, .m4a, .aac, .flac (audio).

---

Embedding cache

Key	Type	Default	Description
cache.enabled	boolean	true	Cache chunk embeddings in SQLite

Prevents re-embedding unchanged text during reindex or transcript updates.

---

Batch indexing

Key	Type	Default	Description
remote.batch.enabled	boolean	false	Enable batch embedding API

Available for gemini, openai, and voyage. OpenAI batch is typically fastest and cheapest for large backfills.

Batch enablement is the only remote batching setting. Concurrency, polling, and timeout behavior are provider-owned.

---

Session memory search

Index session transcripts and surface them via memory_search:

Key	Type	Default	Description
rememberAcrossConversations	boolean	On for personal installs; off with configured DM isolation	Permit private cross-conversation recall
sources	string[]	["memory"]	Add "sessions" to include transcripts

Session indexing is opt-in and runs asynchronously. Results can be slightly stale. Session logs live on disk, so treat filesystem access as the trust boundary. The [session-memory hook](/automation/hooks#session-memory) saves conversation excerpts to `/memory/`, which the `memory` source already indexes. If transcript indexing is also enabled, the same conversation can appear from both `memory` and `sessions`, resulting in overlapping search results and additional embedding work. For hook-only recall, set `sources: ["memory"]` and `rememberAcrossConversations: false`; `sources` alone is insufficient because cross-conversation recall automatically adds `sessions`. For full-transcript recall instead, run `openclaw hooks disable session-memory`. Enable both only when you intentionally want both representations.

Ordinary model-invoked session transcript search obeys tools.sessions.visibility. The default tree visibility exposes the current session, sessions it spawned, and same-agent group sessions watched through ambient group awareness. Other unrelated sessions require agent visibility (or all only when cross-agent recall is also required and agent-to-agent policy allows it).

rememberAcrossConversations does not widen that setting. It supplies a separate runtime-only authorization limited to same-agent private transcripts during the bounded Active Memory pass.

The examples below place these settings under top-level memory.search. You can also apply equivalent settings in a per-agent memory.search override when only one agent should index and search session transcripts.

For same-agent gateway-to-DM recall:

{
  memory: {
    search: {
      experimental: { sessionMemory: true },
      sources: ["memory", "sessions"],
    },
  },
  tools: {
    sessions: { visibility: "agent" },
  },
}

---

SQLite vector acceleration (sqlite-vec)

Key	Type	Default	Description
store.vector.enabled	boolean	true	Use sqlite-vec for vector queries
store.vector.extensionPath	string	bundled	Override sqlite-vec path

When sqlite-vec is unavailable, OpenClaw falls back to in-process cosine similarity automatically.

---

Index storage

Built-in memory indexes live in each agent's OpenClaw SQLite database at agents/<agentId>/agent/openclaw-agent.sqlite.

Key	Type	Default	Description
store.fts.tokenizer	string	unicode61	FTS5 tokenizer (unicode61 or trigram)

---

Citations

memory.citations controls citation visibility for built-in memory results:

Value	Behavior
auto (default)	Include Source: <path#line> when useful
on	Always include the source footer
off	Omit the footer; the path remains available internally

---

Dreaming

Dreaming is configured under plugins.entries.memory-core.config.dreaming, not under memory.search.

Dreaming runs as one scheduled sweep and uses internal light/deep/REM phases as an implementation detail.

For conceptual behavior and slash commands, see Dreaming.

User settings

Key	Type	Default	Description
enabled	boolean	true	Enable or disable dreaming entirely
frequency	string	0 3 * * *	Optional cron cadence for the full dreaming sweep
model	string	default model	Optional Dream Diary subagent model override
phases.deep.maxPromotedSnippetTokens	number	160	Maximum estimated tokens kept from each short-term recall snippet promoted into MEMORY.md; provenance metadata remains visible
phases.deep.maxPriorEntryLossFraction	number	0.25	Reject a consolidation rewrite that removes more than this fraction of prior entries

Example

{
  plugins: {
    entries: {
      "memory-core": {
        subagent: {
          allowModelOverride: true,
          allowedModels: ["anthropic/claude-sonnet-4-6"],
        },
        config: {
          dreaming: {
            enabled: true,
            frequency: "0 3 * * *",
            model: "anthropic/claude-sonnet-4-6",
          },
        },
      },
    },
  },
}

- Dreaming writes machine state to `memory/.dreams/`. - Dreaming writes human-readable narrative output to `DREAMS.md` (or existing `dreams.md`). - Deep consolidation stores the prior `MEMORY.md` in SQLite-backed plugin state and records rewrite counts and highlights in `DREAMS.md`. - Untrusted and system-derived candidates are structurally excluded before consolidation and durable promotion. - `dreaming.model` uses the existing plugin subagent trust gate; set `plugins.entries.memory-core.subagent.allowModelOverride: true` before enabling it. - Dream Diary retries once with the session default model when the configured model is unavailable. Trust or allowlist failures are logged and are not silently retried. - The light/deep/REM phase policy and thresholds are internal behavior, not user-facing config.

Related

• Configuration reference
• Memory overview
• Memory search