Memory search

finds relevant notes from your memory files, even when the wording differs from the original text. It works by indexing memory into small chunks and searching them using embeddings, keywords, or both.

Quick start

If you have a GitHub Copilot subscription, OpenAI, Gemini, Voyage, or Mistral API key configured, memory search works automatically. To set a provider explicitly:

json5
Copy
{
  agents: {
    defaults: {
      memorySearch: {
        provider: "openai", // or "gemini", "local", "ollama", etc.
      },
    },
  },
}

For multi-endpoint setups,

can also be a custom entry, such as , when that provider sets or another embedding adapter owner.

For local embeddings with no API key, set

. Packaged installs retain the native runtime in OpenClaw's managed plugin runtime-deps tree; run if that tree needs repair.

Some OpenAI-compatible embedding endpoints require asymmetric labels such as

for searches and or for indexed chunks. Configure those with and ; see the Memory configuration reference.

Supported providers

How search works

OpenClaw runs two retrieval paths in parallel and merges the results:

mermaid
Copy
flowchart LR
    Q["Query"] --> E["Embedding"]
    Q --> T["Tokenize"]
    E --> VS["Vector Search"]
    T --> BM["BM25 Search"]
    VS --> M["Weighted Merge"]
    BM --> M
    M --> R["Top Results"]

• Vector search finds notes with similar meaning ("gateway host" matches "the machine running OpenClaw").
• BM25 keyword search finds exact matches (IDs, error strings, config keys).

If only one path is available (no embeddings or no FTS), the other runs alone.

When embeddings are unavailable, OpenClaw still uses lexical ranking over FTS results instead of falling back to raw exact-match ordering only. That degraded mode boosts chunks with stronger query-term coverage and relevant file paths, which keeps recall useful even without

or an embedding provider.

Improving search quality

Two optional features help when you have a large note history:

Temporal decay

Old notes gradually lose ranking weight so recent information surfaces first. With the default half-life of 30 days, a note from last month scores at 50% of its original weight. Evergreen files like

are never decayed.

MMR (diversity)

Reduces redundant results. If five notes all mention the same router config, MMR ensures the top results cover different topics instead of repeating.

Enable both

json5
Copy
{
  agents: {
    defaults: {
      memorySearch: {
        query: {
          hybrid: {
            mmr: { enabled: true },
            temporalDecay: { enabled: true },
          },
        },
      },
    },
  },
}

Multimodal memory

With Gemini Embedding 2, you can index images and audio files alongside Markdown. Search queries remain text, but they match against visual and audio content. See the Memory configuration reference for setup.

Session memory search

You can optionally index session transcripts so

can recall earlier conversations. This is opt-in via . See the configuration reference for details.

Troubleshooting

No results? Run

to check the index. If empty, run .

Only keyword matches? Your embedding provider may not be configured. Check

.

Local embeddings time out?

, , and use a longer inline batch timeout by default. If the host is simply slow, set and rerun .

CJK text not found? Rebuild the FTS index with

.

Further reading

• Active Memory -- sub-agent memory for interactive chat sessions
• Memory -- file layout, backends, tools
• Memory configuration reference -- all config knobs

Related

• Memory overview
• Active memory
• Builtin memory engine