Analytics V2

curl --request POST \
  --url https://api.linqalpha.com/v2/analytics/sse \
  --header 'Content-Type: application/json' \
  --header 'X-API-KEY: <api-key>' \
  --data '
{
  "query": "What are the latest insights on AAPL performance?",
  "search_types": {
    "external": [
      {
        "type": "transcript"
      },
      {
        "type": "filing"
      }
    ]
  }
}
'

{
  "event_name": "conversation",
  "data": {
    "conversation_id": "7622b901-93ee-44bb-afb3-d60df76e31bf",
    "message": "Successfully created a conversation."
  }
}

Analytics V2

Agentic analytics with structured source selection, hard/soft filters, and identity fields.

search_types: Choose sources — rms (internal docs), external (transcripts, filings), structured
filters.rms.hard_filters / soft_filters: Hard = exclude non-matching (AND between fields, OR within arrays). Soft = boost relevance only.
organization_id and user_id: Optional (required only for platform API keys)

Event Flow

conversation            ← conversation_id
message                 ← echo of user query
status: start           ← stream begins
+-- AGENTIC LOOP (repeats) ---------+
|  tool_use_block (x1-3)            |
|  tool_result_block (x1-3)         |
|  keepalive (~15s)                 |
|  think (xN)                       |
+------------------------------------+
answer                  ← full answer with [1][2] citations
status: finish          ← stream ends

Retrieving References

After the stream finishes, use conversation_id to get citation sources:

GET /v2/analytics/conversations/{conversation_id}/references

Returns normalized references with enriched metadata. See Analytics V2 References for details.

One question at a time per conversation. A request that arrives while an earlier question with the same conversation_id is still being answered can wait up to about 15 seconds and then end with a status event carrying error whose code is SESSION_BUSY (HTTP status is already 200). The refused question is recorded in the conversation history but is not answered; send it again once the earlier answer completes.

POST

analytics

sse

curl --request POST \
  --url https://api.linqalpha.com/v2/analytics/sse \
  --header 'Content-Type: application/json' \
  --header 'X-API-KEY: <api-key>' \
  --data '
{
  "query": "What are the latest insights on AAPL performance?",
  "search_types": {
    "external": [
      {
        "type": "transcript"
      },
      {
        "type": "filing"
      }
    ]
  }
}
'

{
  "event_name": "conversation",
  "data": {
    "conversation_id": "7622b901-93ee-44bb-afb3-d60df76e31bf",
    "message": "Successfully created a conversation."
  }
}

Referencing Vault documents

To chat over documents you uploaded through the Vault endpoints, pass their rms_document_id (returned by Vault — Confirm Upload) in search_types.rms[].document_ids, and set source to "vault":

{
  "organization_id": "<org uuid>",
  "query": "Summarize the Q3 figures in the uploaded report.",
  "search_types": {
    "rms": [
      { "workspace": "personal", "source": "vault", "document_ids": ["<rms_document_id>"] }
    ]
  }
}

source: "vault" is required for Vault documents. If you omit it, the document is searched generically — it is not retrieved for direct file citation, the answer’s [N] markers won’t resolve, and GET /v2/analytics/conversations/{id}/references returns an empty list (the answer may still look complete, so this fails silently).

document_ids are rms_document_id values from confirm (poll Document Status until Synced first) — not the presign document_id.
workspace must match the workspace the document was uploaded into (the workspace sent to presigned_url / confirm).

Selecting a model

Pass workspace_model to run a request on a specific model:

{
  "organization_id": "<org uuid>",
  "query": "Summarize BNP Paribas' latest credit profile.",
  "workspace_model": "claude-opus-4-7"
}

Omit it and the model configured for your organization is used — the default for almost all requests. Reach for it when a particular query needs more (or less) capability than your default. The model that actually ran is reported on the workspace_model SSE event, so you can always confirm which one served the request:

data: {"event_name": "workspace_model", "data": {"workspace_model": "claude-opus-4-7[1m]"}}

The reported id carries a [1m] suffix marking the 1M-context runtime the model ran under. It is appended to every Claude model, so the reported value will not string-match a bare id you sent.Both forms are accepted on the way in (claude-opus-4-7 and claude-opus-4-7[1m] are equivalent), so you can send back whatever you read. To compare, strip the suffix on either side.

If your organization pins a model, that pin wins: asking for a different model is refused after the stream opens — 200, then a status event carrying error, not a silent success on the pinned model. Omit workspace_model to run on the pin. Read the workspace_model SSE event to see what ran.

A malformed model id — one that does not match the Claude model id format — is rejected with 400 before the stream opens. Passing that check is not the same as being able to run. An id that is well-formed but is not a real model, or is a real model your organization is not configured to use, is refused only after the stream has opened: you receive 200, then a status event carrying error — not a 400. Tell us which models you need and we will enable them.

Authorizations

X-API-KEY

string

header

required

Body

application/json

query

string

required

The analytics query to process.

Example:

"What are the latest insights on AAPL earnings?"

organization_id

string<uuid>

Organization ID

user_id

string<uuid>

User ID

user_email

string<email>

User email

user_name

string

User name

conversation_id

string | null

Existing conversation ID for multi-turn continuation. Omit or null for new conversation.

search_types

object | null

Structured source selection. Controls which document sources to query. Omitting a category means don't search it.

Show child attributes

filters

object | null

Metadata filters. Hard filters exclude non-matching documents (AND between fields, OR within arrays). Soft filters boost relevance without excluding.

Show child attributes

connectors

object[] | null

Customer MCP connectors to enable for this analytics request.

Show child attributes

disabled_connectors

string[] | null

Connector names to disable for this analytics request.

workspace_model

string | null

Optional. LLM model to run this request with, e.g. claude-opus-4-7, claude-opus-5 or claude-opus-5-5. When omitted, the model configured for your organization is used. If your organization pins a model, that pin wins and requesting a different model is refused after the stream opens rather than silently succeeding on the pin; omit this field to run on the pin. The model that actually ran is reported on the workspace_model SSE event — note that the reported id carries a [1m] 1M-context suffix (e.g. claude-opus-4-7[1m]), so it does not string-match a bare id you sent. Both forms are accepted on input, so the reported id can be sent back as-is. A malformed id is rejected with 400 before the stream opens; an id that is well-formed but not available to your organization is refused after the stream opens, as a status event carrying error on a 200.

Pattern: ^claude-(?:opus|sonnet|haiku)-[\w.-]{1,40}(?:\[1m\])?$

Example:

"claude-opus-4-7"

Response

SSE stream of analytics events. Each event is sent as data: {json}\n\n. Events follow the sequence: conversation → message → status:start → [agentic loop] → answer → status:finish. status:keep_alive and keepalive events are sent periodically during long tool execution gaps.

Conversation event

event_name

enum<string>

required

Available options:

conversation

data

object

required

Event-specific data payload

Show child attributes

MCP

Basic

Transcriptions

Evaluations

RMS (On request)

Briefing

Sync Status

Meeting Agent

Analytics V2

Event Flow

Retrieving References

Referencing Vault documents

Selecting a model

Authorizations

Body

Response

​Referencing Vault documents

​Selecting a model

Authorizations

Body

Response

Referencing Vault documents

Selecting a model