GovTribe

Add to vector store MCP tool

Adds authorized GovTribe files or file-bearing entities into a registered vector store hosted by OpenAI so their full text becomes searchable via Search_Vector_Store. Sources and stores must pass GovTribe access checks; arbitrary URLs are not accepted. Returns the current status snapshot quickly; if status is in_progress, call Add_To_Vector_Store again later with the same source items and returned govtribe_vector_store_id.

Examples

When the selected vector store has expired, Add_To_Vector_Store creates a replacement and adds only files requested by the affected add operation. Earlier files are not carried over automatically. The response message discloses replacement on the initial result and later polls. Use the returned new govtribe_vector_store_id for polling and subsequent calls, and explicitly add any earlier files you still need. Wait for successful processing before searching; pending or failed files are not searchable.

Use this when an agent needs semantic retrieval over full file text. A common workflow is a solicitation package review: resolve an opportunity, use Search_Government_Files to confirm the source package, stage the opportunity or selected files with Add_To_Vector_Store, then ask focused questions with Search_Vector_Store.

This is a polling tool. Use Add to vector store MCP response to interpret readiness fields, text-only client behavior, retry instructions, and failures.

In Analyst and Analyst Astra, first call Create_Vector_Store with action: ensure and wait for its status to report a ready store. Pass its current govtribe_vector_store_id to Add; Add does not create the initial Analyst store. If Add reports that the selected store is still being created, call Create with action: status and the original request_id, then retry Add with the same selected items and current store ID. After Add completes, call Create status again to refresh hosted file_search availability. External MCP clients continue to use Add and Search without calling Create.

For complete package analysis, wait until the response says the requested files are ready. If processing is still in progress, poll Add_To_Vector_Store again later with the same source items and the returned govtribe_vector_store_id. A premature external MCP Search_Vector_Store call can return a not-ready message, and partial results should not be treated as a complete package readout.

If the response lists skipped or unsupported spreadsheet-like files, such as pricing schedules, workbooks, CSV/TSV files, CLIN tables, budget templates, or staffing matrices, the vector store only covers the supported files. Use Add_To_Container and hosted shell extraction for those skipped files before making complete source-package claims.

External MCP: add a solicitation package to a new vector store

{
  "tool": "Search_Federal_Contract_Opportunities",
  "arguments": {
    "query": "facility modernization statement of work",
    "search_mode": "semantic",
    "fields_to_return": [
      "govtribe_id",
      "name",
      "government_files"
    ]
  }
}
{
  "tool": "Add_To_Vector_Store",
  "arguments": {
    "items": [
      {
        "govtribe_type": "federal_contract_opportunity",
        "govtribe_id": "<federal_contract_opportunity_govtribe_id>"
      }
    ]
  }
}

Check an existing vector store while files are still processing

{
  "tool": "Add_To_Vector_Store",
  "arguments": {
    "govtribe_vector_store_id": "<govtribe_vector_store_id>",
    "items": [
      {
        "govtribe_type": "government_file",
        "govtribe_id": "<government_file_govtribe_id>"
      }
    ]
  }
}

Payload patterns

Call the tool again when the response says processing is still in_progress. Keep the same source items, and include the returned govtribe_vector_store_id on follow-up calls. Once the response says the requested supported files are ready, search the returned govtribe_vector_store_id with Search_Vector_Store. If the response includes skipped files, unsupported file guidance, or failure guidance, use it before choosing another file, falling back to metadata/snippet search, or staging skipped files with Add_To_Container.

Stage a pursuit file set for later semantic retrieval

{
  "tool": "Add_To_Vector_Store",
  "arguments": {
    "items": [
      {
        "govtribe_type": "pursuit",
        "govtribe_id": "<pursuit_govtribe_id>"
      },
      {
        "govtribe_type": "user_file",
        "govtribe_id": "<capture_notes_user_file_govtribe_id>"
      }
    ]
  }
}