Discussion about this post

User's avatar
xiq's avatar

Technical part of the community archive post:

# Building

How does the API work?

- Upload through our website

- The API is read-only

- Access via curl requests or Supabase client libraries

- See our README for more details and examples

Will you offer full dumps of the data? Yes.

Will the API always be free? It'll remain free as long as we can sustainably run it. While we reserve the right to discontinue the project, we'll make every effort to distribute full data dumps to interested parties.

Are you storing pictures too? Not yet due to storage costs, but we'd love to in the future.

What are some ideas for projects using this data?

- Semantic search over tweets;

- Self-knowledge applications like analyzing your trends over time and summaries of your thoughts on topics you're usually talking about;

- Digital anthropology and sociology research ("what are the origins of this idea / project / movement?", "how did these people start interacting?")

- Cross-archive search like "which of my friends are talking about this topic?"

- Producing stuff like books from your best tweets;

- Migrating to alternative networks like Bluesky could be made easier if the canon of important tweets is easy to migrate.

- If someone is active enough, you could produce a "digital twin" from their archives by e.g. fine-tuning an LLM on their tweets, or just selecting the best and putting them in context.

# Logistics

Will people have to keep re-uploading their archives to stay up to date? Yes, for now. We're focused on preserving historical data. Future ideas:

- Organize yearly "upload parties"

- Develop a browser extension for automated updates

How hard is it to get tweets otherwise? Twitter severely limits access. Current rates: $5000 for 1M tweets (far fewer than we aim to preserve, at a much higher cost).

Expand full comment
Tasshin Fogleman's avatar

doing the lord's work, bless ❤️

Expand full comment
2 more comments...

No posts