OSIA

Open Source Instagram Aggregator

OSIA is a service that maintains a copy of your Instagram feed and exposes it with a simple REST API. It is especially convenient for displaying instagram posts on a website.

The application is composed of an aggregator, which periodically checks new posts and saves them on a local database, and an HTTP server, which serves posts from the local database.

Why do I need that ?

Due to past abuses (Cambridge Analytica user data scandal), Facebook restricted its APIs, including the Instagram one. Even if an Instagram account is public, fetching its list of posts now requires a valid token that must be renewed at regular interval. Sadly (for developers, but probably not for Instagram users), it prevents a website from directly fetching posts from Instagram. Therefore, a third service is needed to collect posts and serve them.

Authentication

OSIA uses the Instagram basic display API. To fetch posts from a public Instagram account you need to have a valid token. This token can be generated with the Instagram token generator. Prior to that, you will need to create a Facebook app with the basic instagram API product. Follow instructions from the official documentation: https://developers.facebook.com/docs/instagram-basic-display-api/getting-started?locale=en_US#step-1--create-a-facebook-app.

A token has a limited value, but can be renewed. To keep the token valid, OSIA renews it every time it makes an API call to check the latest posts.

Requirement

You can use the existing binaries from the releases section, which are completely self-contained. No additional installation is needed. If you have Go installed (>=1.18), you can also compile the app yourself. From the root folder:

# build a binary
go build .
# build a binary and moves it to $GOBIN
go install .

Usage

The Instagram token must be passed as an environment variable, and settings can be passed via the CLI. To see the options, use -h. For example:

INSTAGRAM_TOKEN=XXX ./OSIA-linux-amd64-a.b.c \
  --interval 30m \
  --dbfilepath data/osia.db \
  --imagesfolder images/ \
  --listen 0.0.0.0:3333

The app can be stopped with Ctrl + C. To prevent a full download, it can be re-started with the same database and images folder.

Read your posts

An HTTP server is bootstrapped at the provided (or default) listen address. It serves the list of medias at the http://<listen>/api/medias endpoint. By default, the endpoint returns a maximum of 12 medias, sorted by timestamp. Recall that your endpoint is likely to be public, you don't want to expose too much data.

It is possible to retrieve less than 12 medias by specifying a count= URL parameter:

# Returns the last 8 posts
http://0.0.0.0:3333/api/medias?count=8

A post has the following attributes:

{
  id:
  caption:
  media_type:
  media_url:
  permalink:
  username:
  timestamp:
}

Images

Due to Instagram security restrictions, images hosted by Instagram cannot be displayed on external websites. Consequentely, a simple <img src={media_url}/> tag would not work. To get around that, images are saved locally to the provided (or default) imagesfolder and served at the http://<listen>/images/<post id>.jpg endpoint. "media id" corresponds to the id of the post.

Use docker

You can quickly use OSIA with docker. The following fetches the latest images from the github registry and runs OSIA. After that you can open your browser on localhost:3333/api/medias and see your posts. Data are also persisted on a data/ folder that you can re-use between runs:

docker pull ghcr.io/nkcr/osia-deploy:latest
docker run -e INSTAGRAM_TOKEN=XXX -p 3333:3333 -v $(pwd)/data:/data ghcr.io/nkcr/osia-deploy

Deployment

It is recommended to use systemd and a reverse proxy, such as nginx. Place osia.service in /etc/systemd/system/. Replace the INSTAGRAM_TOKEN with yours.

Create the following folders:

sudo mkkdir /opt/osia/bin

Download and move the OSIA binary to /opt/osia/bin. Then create an osia user+group, and update the ownership of the osia folder:

wget https://github.com/nkcr/OSIA/releases/download/va.b.c/osia-linux-amd64-va_b_c /tmp
sudo mv /tmp/osia-linux-amd64-va_b_c /opts/osia/bin/osia

sudo useradd -r osia
sudo chown -R osia:osia /opt/osia

Optionally, configure logrotate by copying osia.logrotate to /etc/logrotate.d/. You should also configure a reverse proxy, see osia.nginx.

Finally, start the osia service:

sudo systemctl enable --now osia.service

You should then see the starting logs with sudo service osia status.

Practical use

Here is a minimal example using react to render a list of posts using OSIA's REST API.

const Instagram = () => {
  const [posts, setPosts] = useState([])

  useEffect(() => {
    fetch(`${ENDPOINT}/api/medias?count=6`)
      .then(response => response.json())
      .then(resultData => {
        setPosts(resultData)
      })
  }, [])

  return (
    <div>
      {posts.map((post) => (
        <Post key={post.id} post={post}/>
      ))}
    </div>
  )
}

const Post = ({ post }) => {
  return (
      <section>
        <a href={post.permalink}>
          <img src={`${ENDPOINT}/images/${post.id}.jpg`}/>
        </a>
        <div>
          <p dangerouslySetInnerHTML={{ __html: post.caption }}></p>
          <p>{post.timestamp}</p>
        </div>
      </section>
  )
}

Dependencies

• buntdb a great key-value store for storing the posts (MIT license)
• go-flags for argument parsing (BSD-3-Clause license)
• zerolog for logging (MIT license)
• testify for unit testing (MIT license)