Connector Types

Connectors allow you to query your data in Evidence Studio. You connect to your data source once, then use that data across all your reports. Evidence supports five types of connectors: Databases, Object Storage, Data Lakes, Flat Files, and Applications.

---

​

Databases

Database connectors connect to your existing data warehouse or database. Evidence extracts the data, and you can define a schedule for sync. How it works:

1. You provide connection credentials to your database
2. Evidence runs queries against your database to extract data
3. Data is cached in Evidence’s query engine in your chosen region
4. Reports query the cached data for fast performance

You choose where your extracted data is stored. See available regions. Best for:

• Existing data warehouses (Snowflake, BigQuery, Redshift, etc.)
• Data that’s already modeled and ready to report on
• Teams who want Evidence to manage the data pipeline

Supported databases:

• Snowflake
• BigQuery
• PostgreSQL
• RDS PostgreSQL
• Redshift
• MotherDuck
• Athena
• MySQL
• SQL Server
• Azure SQL Database
• Azure Postgres
• Databricks

We are continuously adding support for new databases. Request one here.

---

​

Object Storage

Object storage connectors read Parquet files directly from cloud storage buckets that you control. Evidence queries the data in place using ClickHouse’s S3 table engine. In addition to the below named providers, any S3 compatible storage is supported (via the “custom S3” connector). How it works:

1. You store Parquet files in your own bucket (S3, GCS, Azure, etc.)
2. You provide bucket credentials to Evidence
3. Evidence queries the Parquet files directly — no data is copied

Key characteristics:

• Instant data availability: When you update a Parquet file in your bucket, the new data is immediately available in Evidence—no sync or publish required
• Schema changes require publish: Adding or removing columns requires clicking Publish or setting up a refresh schedule
• Data stays in your infrastructure: Your data never leaves your bucket, Evidence queries it in place
• S3-compatible: Any storage provider with an S3-compatible API works (AWS S3, GCS, Azure Blob, Cloudflare R2, MinIO, Backblaze B2, etc.)

Best for:

• Data residency or sovereignty requirements
• Teams who manage their own data pipelines and produce Parquet files
• High-frequency data updates where you control the source files
• Keeping data within your own infrastructure

Supported providers:

• AWS S3
• Google Cloud Storage
• Azure Blob Storage
• Cloudflare R2
• Backblaze B2
• Custom S3-compatible provider

---

​

Data Lakes

Data lake connectors connect to open table formats that provide ACID transactions, schema evolution, and time travel capabilities on top of object storage. How it works:

1. You provide credentials to your data lake catalog or storage
2. Evidence reads the table metadata and data files
3. Data is queried with full support for the table format’s features

Best for:

• Teams using modern lakehouse architectures
• Large-scale data with complex partitioning
• Scenarios requiring time travel or schema evolution

Supported data lakes:

• Iceberg
• Delta Lake

---

​

Applications

Application connectors sync data from SaaS applications and APIs. Evidence uses Fivetran to extract and normalize data from these sources. How it works:

1. You authenticate with the application (OAuth or API key)
2. Evidence syncs data on a configurable schedule
3. Data is normalized and stored in Evidence’s query engine

Best for:

• Product analytics and operational reporting
• Combining SaaS data with your data warehouse
• Teams who want turnkey integrations without building pipelines

Supported applications:

• Linear
• HubSpot
• Stripe
• GitHub
• Attio
• PostHog
• Google Ads
• Facebook Ads
• Twilio

We are continuously adding support for new applications. Request one here.

---

​

Flat Files

Upload flat files directly to Evidence for quick data analysis without setting up external connections. How it works:

1. Upload a file directly through the Evidence interface
2. Evidence processes and stores the data
3. Query the data immediately in your reports

Supported formats:

• CSV
• Parquet
• JSONL
• Excel

Best for:

• Quick prototyping and ad-hoc analysis
• Small datasets that don’t require a database
• Getting started with Evidence before connecting external sources

See Flat Files for details.

---

​

Comparison

---

​

Storage Regions

Evidence supports the following storage regions for extracted data.