If you've spent any time building AI agents, you know how the workflow actually goes.
The model reasoning is solid, the orchestration layer is wired up, and the output parsing is handled. Then comes the data retrieval step, and suddenly your team is deep in provider's documentation, mapping field schemas, and debugging why a query that works in the sandbox returns zero results in production.
This is the friction that kills momentum in AI development.
Not the model or the infrastructure but the data integration layer.
Coresignal's Agentic Search API is built to eliminate this issue.
What the Agentic Search API actually does
With our new Agentic Search API, you can now integrate this API into your product, use it to describe the data you need in plain language, and get structured B2B data back in your own app or dashboard.
No DSL syntax, schema mapping, or translation layer to maintain on your end.
Your end-users can send a prompt like "Senior Product Managers at fintech companies in London with five or more years of experience" and our API handles the rest: it interprets the intent, translates it into a precise query against Coresignal's database, and returns structured, field-rich results ready for downstream use in your pipeline.
So, how does Agentic Search API work?
- You integrate the Agentic Search API into your product
- Your end-user describes what they need in plain language
- The API translates their intent into an ElasticSearch DSL query
- Structured results are returned directly into your product
- Optional: Request the generated query itself to inspect, modify or reuse it
The results are delivered using Coresignal's multi-source datasets with 839M employee profiles, all ethically collected from public sources. The dataset contains 300+ data fields from multiple integrated and deduplicated sources. This level of data quality and data transparency are not offered by competing natural language data APIs. Read more about the feature in our documentation.
In the upcoming months, we will also start providing access to our multi-source company and jobs datasets, too.
Built for speed and scale
The /fast endpoint is optimized for speed and cost. It accepts a natural-language prompt and an entity type (company, employee, or job), translates the prompt using a simplified schema, and returns results within seconds.
This is the endpoint for high-volume programmatic use for:
- Powering AI sourcing agents
- Embedding search into customer-facing products
- Running automated enrichment pipelines where thousands of queries need to execute reliably and cheaply
The endpoint returns up to 100 data records for preview or the Elasticsearch DSL query itself, depending on your preferred output format.
If you want data, you get data. If you want to validate the generated query first, you get that instead.
Why this matters for AI product builders
Teams building AI agents invest weeks learning the query language, mapping schemas, and handling edge cases before they retrieve a single meaningful result. The Agentic Search API compresses that into a single API call.
For teams building on top of Coresignal’s multi-source data, this changes what's possible at the product level:
- Developers building agentic pipelines get a B2B data retrieval step that fits natively into an LLM-powered workflow.
- Recruitment and HR tech platforms can embed a natural language candidate search experience for non-tech users.
- Sales tech and GTM platforms can let sales reps describe their ideal prospect in a sentence rather than navigating filter menus.
- Research teams can run exploratory queries against professional data without needing to understand Elasticsearch querying.
How to test this new API endpoint?
Unlike general-purpose natural language search tools, the Agentic Search API is built specifically for B2B entities. The system understands seniority levels, industry taxonomies, headcount ranges, company categories, and the relationship between fields in a professional data context.
This translates into higher precision for B2B use cases compared to systems trained on general web content.
Test Agentic Search API
Interested in trying out this new API endpoint?
- Get your API key
You’ll need the API key to authenticate your requests. If you don’t have an API key yet, create an account on our self-service, then copy your API key from the homepage of our dashboard.
- Try the sample request with the new endpoint
We’ve prepared a quick example prompt to show what this new API can do. Simply add an API key, copy this prompt, and call the new endpoint.
curl -X POST "https://api.coresignal.com/cdapi/v2/agentic_search/fast" \
-H "Content-Type: application/json" \
-H "apikey: <YOUR_API_KEY>" \
-d '{
"prompt": "Find founders or co-founders who have previously held an engineering role at Google, Meta, Apple, Amazon, or Microsoft.",
"return_data": true,
"entity": "employee",
"limit": 20
}'
Good luck! If you have any questions, check out our documentation or connect with our team. We’re here to help you evaluate your data needs, walk through available options, and identify the right approach for your use case.
