> ## Documentation Index
> Fetch the complete documentation index at: https://developers.techwolf.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Key concepts

Datasource integrations play a crucial role in populating the TechWolf
SkillEngine API with data from different systems. This document outlines our
**vision** for datasource integrations, explains the differences between
file‑based and API‑based integrations, and clarifies what is **possible** with
our current datasource integration offering.

To build your own datasource integration, see the
[Datasource Integration Guide](/integrations/datasource-integrations/how-to-datasource-integration).
It provides an overview of how TechWolf integrates datasources and explains the
challenges involved.

## Vision

At TechWolf, our vision for datasource integrations is to make unstructured and
semi-structured data usable for skill inference, optimizing for
**time-to-value**, and **minimal back-and-forth communication**, both on
TechWolf's side and the customer's. We aim to:

* Accept the variability of real-world enterprise data.
* Enable low-friction, high-trust integrations, especially for customers
  without resources to build their own integrations.
* Build highly specialized connectors for common use cases, and standardize
  platforms through file‑based integrations to increase coverage.
* Make it easy to onboard, maintain, and evolve integrations across different
  data types.
* Ensure integrations can also be built by customers; we use our own API to
  ingest data.

## File-based vs API-based integrations

File‑based integrations ingest data from files exported by the source system,
either directly or via other export methods (e.g., API exports).

Examples of file-based integrations:

* Workday EIB exports
* SAP Integration Center or Report Center export integrations
* Jira ticket exports to CSV

API‑based integrations pull data from a source API or react to webhooks, and
submit that data to a destination API.

Examples of API‑based integrations:

* Workday SOAP API syncs
* Consuming Jira webhooks
* Consuming GitHub webhooks

A comparative table:

|                      | File-based                                                                      | API-based                                                                 |
| -------------------- | ------------------------------------------------------------------------------- | ------------------------------------------------------------------------- |
| Destination          | TechWolf API                                                                    | TechWolf API                                                              |
| Format               | CSV, Parquet, etc.                                                              | JSON, XML, etc.                                                           |
| Ingestion strategy   | Work with files                                                                 | Consume API or webhooks                                                   |
| Frequency            | Daily, weekly, etc.                                                             | Real-time, or also daily, weekly, etc.                                    |
| Volume               | Limits apply to destination API                                                 | Both source and destination API limits apply                              |
| Eventual consistency | Easier: you can combine files to produce complete datasets                      | Harder: you must track failures somewhere (in files or other storage)     |
| Maintenance          | Keep file formats consistent                                                    | Keep up with development of the source API                                |
| Observability        | Files clearly indicate the data to be synced                                    | API logs and monitoring track what data leaves the source system          |
| Reusability          | Files are always required, but loading functionality is reusable across sources | Easy to build a common foundation, but each source must be re‑implemented |

## What is possible today?

### Standardized Integrations

See our [Integrations Overview](/integrations/overview) of the standardized
integrations we offer, such as Workday or SAP. We will continue building these
standardized integrations, to provide the best onboarding and integration
experience for our customers.

### Other platforms

While we have standardized integrations for specific platforms that provide
optimized data processing and seamless connectivity, TechWolf's flexible
file-based integration approach allows us to ingest data from virtually any
business system, for example:

* Your company's custom learning platform
* Salesforce data
* ServiceNow
* ...

If there is a file for it, we can ingest it into the TechWolf API.

### Supported Data Types

To see which types of data we currently support, we refer to the
[API Documentation](/reference). If your data can map to the API, we can ingest
it through file-based integrations.

For example, to see which fields of courses can be ingested, see the
[Courses API](/reference/latest/Courses/CourseCRUD/overview).

### Supported File Types

We currently only support `.csv` files for standard integrations. If you need to
ingest other file types, please reach out to us.