kipp

Getting started

The easiest way to get started with kipp is by using the image published to Docker Hub. The service is then available simply by running:

docker pull uhthomas/kipp
docker run uhthomas/kipp

Databases

Databases can be configured using the --database flag. The flag requires the input be parsable as a URL. See the url.Parse docs for more info.

Badger

Badger is a fast, embedded database which is great for single instances.

SQL

Kipp uses a generic SQL driver, but currently only loads:

• PostgreSQL

As long as a database supports Go's sql package, it can be used. Please file an issue for requests.

File systems

File systems can be configured using the --filesystem flag. The flag requires the input be parsable as a URL. See the url.Parse docs for more info.

Local (your local file system)

The local filesystem does not require any special formatting, and can be used like a regular path such

--filesystem /path/to/files

AWS S3

AWS S3 requires the s3 scheme, and has the following syntax:

--filesystem s3://some-token:some-secret@some-region/some-bucket?endpoint=some-endpoint.

The region and bucket are required.

The user info section is optional, if present, will create new static credentials. Otherwise, the default AWS SDK credentials will be used.

The endpoint is optional, and will use the default AWS endpoint if not present. This is useful for using S3-compatible services such as:

• Google Cloud Storage - storage.googleapis.com
• Linode Object Storage - linodeobjects.com
• Backblaze B2 - backblazeb2.com
• DigitalOcean Spaces - digitaloceanspaces.com
• ... etc

Policy

Required actions:

• s3:DeleteObject
• s3:GetObject
• s3:PutObject

This is subject to change in future as more features are added.

Building from source

Kipp builds, tests and compiles using Bazel. To run/build locally with bazel:

git clone git@github.com:uhthomas/kipp
cd kipp
bazel run //cmd/kipp

API

Kipp has two main components; uploading files and downloading files. Files can be uploaded by POSTing a multipart form to the / endpoint like so:

curl https://kipp.6f.io -F file="some content"

The service will then respond with a 302 (See Other) status and the location of the file. It will also write the location to the response body.

Kipp also serves all files located in the web directory by default, but can either be disabled or changed to a different location.