Integration

Vald integration

Vald is a self-hosted distributed approximate nearest-neighbour vector search engine that runs on Kubernetes and exposes a REST gateway beside its gRPC one. This integration covers that gateway: insert, update, upsert and remove vectors one at a time or in a batch, remove or re-stamp vectors by timestamp, search by query vector or by the id of a vector already indexed, run the same searches in linear (exact) mode, read a stored vector, its timestamp or whether an id exists, push objects through a configured filter target so the ingress filter vectorises them before insert, update, upsert or search, read index information, per-agent index detail, index properties and index statistics, and flush the whole index. The host is part of the credential because every Vald deployment is its own cluster; the REST gateway carries no authentication of its own, so protect it at the ingress.

Connect Vald How integrations work

Ask it in chat

Say what should happen.
The workflow builds itself.

When this happens, do that in Vald.Built from the 35 Vald tools below, on the connection you add once.

What you can do with Vald

Each of these is one operation on one endpoint, available both as a workflow step and as a tool an AI agent can call on its own. Nothing bundles several operations into one action, so what a run did stays readable afterwards.

Most endpoints run on the free plan. A few, mainly the ones that publish on your behalf, need a paid plan.

Multiple, 12 operations

insert_objectsInsert objectsPOSTInserts several objects, letting the ingress filter turn each into a vector.
insert_vectorsInsert vectorsPOSTInserts several new vectors in one call.
linear_search_by_idsLinear search by idsPOSTRuns several exact linear searches, each keyed on an indexed vector id.
linear_search_by_vectorsLinear search by vectorsPOSTRuns several exact linear searches, each with its own query vector.
remove_vectorsRemove vectorsPOSTRemoves several indexed vectors in one call.
search_by_idsSearch by idsPOSTRuns several approximate searches, each keyed on an indexed vector id.
search_by_objectsSearch by objectsPOSTRuns several searches from raw objects, each vectorised by the ingress filter.
search_by_vectorsSearch by vectorsPOSTRuns several approximate searches, each with its own query vector.
update_objectsUpdate objectsPOSTReplaces several stored vectors with ones the ingress filter derives from objects.
update_vectorsUpdate vectorsPOSTReplaces several stored vectors in one call.
upsert_objectsUpsert objectsPOSTInserts or replaces several vectors the ingress filter derives from objects.
upsert_vectorsUpsert vectorsPOSTInserts or replaces several vectors in one call.

Object, 5 operations

get_vectorGet vectorGETReads an indexed vector back, with the timestamp it was stored with.
insert_objectInsert objectPOSTInserts one object, letting the ingress filter turn it into a vector.
search_by_objectSearch by objectPOSTSearches with a raw object, letting the ingress filter turn it into the query vector.
update_objectUpdate objectPOSTReplaces a stored vector with one the ingress filter derives from an object.
upsert_objectUpsert objectPOSTInserts or replaces a vector the ingress filter derives from an object.

Detail, 2 operations

get_index_detailGet index detailGETReads per-agent index counts and replica placement.
get_index_statistics_detailGet index statistics detailGETReads graph statistics per agent.

Id, 2 operations

linear_search_by_idLinear search by idPOSTSearches for the nearest neighbours of an already indexed vector, by scanning every vector.
search_by_idSearch by idPOSTSearches the approximate index for the nearest neighbours of an already indexed vector.

Timestamp, 2 operations

remove_vectors_by_timestampRemove vectors by timestampPOSTRemoves indexed vectors selected by their timestamp.
update_vector_timestampUpdate vector timestampPOSTWrites a new timestamp on an indexed vector without touching its values.

Exist, 1 operation

vector_existsVector existsGETChecks whether a vector is indexed under the given id.

Flush, 1 operation

flush_indexFlush indexDELETERemoves every vector from the Vald index.

Info, 1 operation

get_index_infoGet index infoGETReads the index counters of the whole cluster.

Insert, 1 operation

insert_vectorInsert vectorPOSTInserts one new vector under a given id.

Linearsearch, 1 operation

linear_search_by_vectorLinear search by vectorPOSTSearches for the nearest neighbours of a query vector by scanning every vector.

Meta, 1 operation

get_vector_timestampGet vector timestampGETReads the timestamp stored with an indexed vector.

Property, 1 operation

get_index_propertyGet index propertyGETReads the graph index properties of every agent.

Remove, 1 operation

remove_vectorRemove vectorPOSTRemoves one indexed vector, named by id.

Search, 1 operation

search_by_vectorSearch by vectorPOSTSearches the approximate index for the nearest neighbours of a query vector.

Statistic, 1 operation

get_index_statisticsGet index statisticsGETReads graph statistics aggregated over the whole index.

Update, 1 operation

update_vectorUpdate vectorPOSTReplaces the vector stored under a given id.

Upsert, 1 operation

upsert_vectorUpsert vectorPOSTInserts a vector, or replaces it when the id is already indexed.

Use Vald from Claude Code, Cursor or Codex

Connect Vald in LiveContext, create an API key in Settings, then add the LiveContext MCP server to your client. Its Vald operations become tools the client can call, next to every other integration you connected.

Claude Code

claude mcp add --transport http livecontext https://livecontext.ai/mcp --header "X-API-Key: YOUR_API_KEY"

Codex (~/.codex/config.toml)

[mcp_servers.livecontext]
url = "https://livecontext.ai/mcp"

[mcp_servers.livecontext.http_headers]
"X-API-Key" = "YOUR_API_KEY"

Cursor, Claude Desktop and other JSON configs

{
  "mcpServers": {
    "livecontext": {
      "type": "http",
      "url": "https://livecontext.ai/mcp",
      "headers": {
        "X-API-Key": "YOUR_API_KEY"
      }
    }
  }
}

Set up the MCP server step by step

How to connect Vald

Create your workspace
Sign up free. Nothing to install.
Connect Vald once
Add your Vald connection in Settings. Every tool from this integration reuses it, and secrets are encrypted at rest and visible only to you.
Describe the job in chat
The workflow builds itself in front of you and picks the endpoints it needs. Drag the steps yourself on the canvas whenever you would rather be explicit.