Screen for funding predictions

Filter prediction records by score, then inspect their dates and timing.

Before you begin

This example was verified with Predictions and Insights access. Your API key must include the fields and collections requested below. Check your API package if an endpoint or field is unavailable. Complete your first API request to set CRUNCHBASE_API_KEY in bash or zsh. Commands below include response headers; look for HTTP 200, then read the JSON body.

Persona: Wealth and asset managers

Use case: Screen for funding predictions

Primary APIs: Funding Prediction Search

Goal: Screen funding prediction records with an overall probability score of at least 0.90. Inspect their generation dates and timing values before prioritizing companies for research. This workflow screens potential fundraising, not IPOs or acquisitions.

Before interpreting the results, review prediction scores and timing. See API Packages Overview for the reference matching your licensed access.

Endpoint

https://api.crunchbase.com/v4/data/searches/funding_predictions

API Reference & Documentation

Query

Returns funding prediction records with probability_score >= 0.90. The overall score has no fixed time horizon. Timing values describe separate month intervals relative to generated_on; do not multiply them by the overall score or treat them as its mathematical decomposition.

Key Entities:

field_idWhat this field doesoperator_id (these are the compatible operators per field)
organizationUsed to identify the company receiving fundingblank;contains;eq;includes;not_contains;not_eq;not_includes;starts
probability_scoreThe predicted probability that this Organization will raise a funding roundbetween;blank;eq;gt;gte;lt;lte;not_eq
probability_score_timeseriesModel timing values grouped into future month intervalsNot compatible with operators

Run the request

curl --include --request POST \
  --url 'https://api.crunchbase.com/v4/data/searches/funding_predictions' \
  --header "X-cb-user-key: $CRUNCHBASE_API_KEY" \
  --header 'Accept: application/json' \
  --header 'Content-Type: application/json' \
  --data '{
  "field_ids": [
    "organization",
    "probability_score",
    "probability_score_timeseries",
    "generated_on"
  ],
  "query": [
    {
      "field_id": "probability_score",
      "operator_id": "gte",
      "type": "predicate",
      "values": [
        0.9
      ]
    }
  ],
  "limit": 3
}'

Response

Response examples below are historical, shortened JSON samples; search samples show one entity and card samples show at most one record per card. Other fields or records may be omitted. Counts and values change; count is the total matching records, while entities contains the returned page.

The shortened response shows one verified funding prediction, including generated_on. Its month-interval values are separate model outputs; the overall score is not a probability for the next five months.

{
  "count": 244791,
  "entities": [
    {
      "uuid": "000059c4-ad07-4572-99c6-994ed2439541",
      "properties": {
        "organization": {
          "permalink": "justiguide",
          "image_id": "7dbc0f9b6b0b463a9bfbd22ebfcc56b7",
          "uuid": "fc5bf38a-0c05-4891-a0cb-5f864d759612",
          "entity_def_id": "organization",
          "value": "JustiGuide"
        },
        "generated_on": "2026-02-21",
        "identifier": {
          "image_id": "bdf23c1234064c60a6dd0b2a878d2d7c",
          "uuid": "000059c4-ad07-4572-99c6-994ed2439541",
          "entity_def_id": "funding_prediction",
          "value": "Funding Prediction 2026-02-21"
        },
        "probability_score": 0.9307,
        "probability_score_timeseries": {
          "months_06_to_11": 4.771613690175369e-08,
          "months_00_to_05": 0.9946821094759647,
          "months_12_to_24": 2.620306727516924e-05,
          "months_24_plus": 0.005291639740623204
        }
      }
    }
  ]
}

Outcome

The response count reports matching funding prediction records. A request returns one page of prediction records, which may include multiple predictions for the same organization. Paginate through the results and group by organization UUID when building a company list. Because this search filters on score before returning records, the newest matching prediction may not be the organization’s newest prediction overall. For a current-company screen, retrieve all predictions for each candidate organization, select the greatest generated_on, then apply the score threshold; see the funding-prediction lookup workflow. Funding predictions describe potential fundraising; they do not predict an acquisition or IPO.

Lists like these can be fed directly into sourcing, coverage models, CRM workflows, or alerting systems, enabling your team to prioritize high-momentum companies, allocate resources more strategically, and engage decision-makers before capital events become widely known.

Check your result

Check the organization UUID and prediction generation date before using a score. An empty result is not a zero probability. Interpret scores and timing using the prediction definitions.

Search requests here use limit: 3 for a small first page. For a complete result set, follow search pagination while preserving the filters and ordering. For request failures, use the shared request and retry guidance.


Did this page help you?