API Playgrounds
API Playgrounds offer a hands-on environment to learn how to search, download, and enrich data using Coresignal's APIs. They’re designed to support onboarding, simplify integration, and help you test requests with minimal setup.
You can use the playgrounds to:
Generate API queries using natural language input
View and adjust the full request structure
Preview sample responses
Explore different data processing levels: Multi-source, Clean, or Base
Available playgrounds:
Each playground includes sample prompts to help you get started. You can edit both the AI-generated query and the original prompt at any time. In the request window, you can:
Add, remove, or update filters
Modify record IDs
Preview different data samples
API playgrounds environment
API playgrounds use entered natural language prompts, which are transformed into search and collect API queries. It simplifies the exploration process and makes it easy to generate custom queries, helping you avoid errors.
Once a request is generated, the tool provides:
The complete search and collect API requests
The total number of matching records in our database
A sample record for quick check
Credits
Generating and sending a request in the API Playgrounds uses credits, following the same credit logic as our general Self-service approach. Find information about credits in Quickstart tutorial.
Prompt information
Rules
Follow the structure of the prompt examples provided in the playground.
Keep it short (5-500 characters) and refine your search through multiple prompts instead of overcomplicating a single prompt.
Use precise criteria instead of broader adjectives. For example, instead of “biggest” or “top”, indicate a company size, revenue range, exact job role, required experience duration, etc.
Aim to formulate prompts that resemble real data fields and values.
The original prompt can be editable – refine and rerun as needed to fine-tune the results.
Prompt errors
When using the API playgrounds, you may run into one of the following errors:
Failed to submit the request
This error message appears when the system cannot process your query as written.
Here’s how you can try to adjust your prompt:
Remove requests for private data. The most common reason for an error is an off-topic request. The AI cannot search for personal contact details, so you should remove any mention of emails or phone numbers from your prompt.
Ensure the prompt is well-formed. A query can fail if it has a malformed or unsupported structure. The best way to fix this is to use one of the prompt examples as a guide to ensure your request is structured correctly.
Avoid using test inputs. Prompts that consist of simple test inputs, typos, or random characters (e.g., “ddkdopqkwd”) are considered unsupported and will cause an error.
Make sure your prompt uses clear, descriptive language.
Insufficient credits
You’ll see this error message if you don’t have enough credits to generate or update a query.
Troubleshooting
No results
If there are no results in our database, the post response field will display 0. However, if the API playground fails to generate a data collection request, it indicates that the API couldn’t find any results based on the search filters.
However, that doesn’t necessarily mean that we don’t have such data. Try to modify your prompt following the best practices:
Broaden your search criteria. Using too many or overly strict filters is a primary cause of empty results. Try using broader criteria first, then narrow the search down as needed.
Check for typos and spelling errors. Requests often fail due to simple misspellings in the prompt. Ensure all names and terms are spelled correctly.
Verify that the entity exists. The company or employee you are looking for may be rare, new, or not available in our dataset.
If you have followed the advice above, it is likely that the data you’re looking for is not available in our database.
Unsuccessful requests
If you run into an error when trying to generate a search query, your request was unsuccessful. Try sending the request again. You won’t be charged for unsuccessful requests.
Last updated
Was this helpful?