Chat with your newsletters.Chat with your newsletters.Chat with your newsletters.Chat with your newsletters.

You subscribe to a dozen newsletters. You read three of them, skim two, ignore the rest. All of them contain something you'd want to find later, and none of them are searchable. This is the database version of your subscriptions.

One inbox on Inbin, 5 newsletter or alert subscriptions pointed at it. Every email that lands gets parsed against your schema.

One PUT to /v1/schemas. Inbin extracts these fields from every future email. Hallucination guard drops anything the model invented.

{
  "extract": {
    "posts": {
      "type": "array",
      "items": {
        "headline":  { "type": "string", "required": true },
        "summary":   { "type": "string" },
        "url":       { "type": "string" },
        "entities": {
          "type": "array",
          "items": { "type": "string" },
          "description": "Companies, people, products mentioned. LLM extracts."
        },
        "topics":    { "type": "array", "items": { "type": "string" } }
      }
    }
  }
}

1. 01
   Create the inbox

   POST /v1/inboxes returns a permanent forwarding address like abc123@in.inbin.dev. Use it everywhere you subscribe.
2. 02
   Point the sources at it

   Sign up for each source with that address, or set a Gmail filter forwarding TLDR, Money Stuff, etc. to the inbox.
3. 03
   Declare the schema

   Paste the JSON above. Inbin creates version 1, tracks all future versions, replays extractions if you change the shape.
4. 04
   Consume the events

   Every parsed email hits your webhook as HMAC-signed JSON. Or read them via GET /v1/events. Or query them from an agent via MCP.

vs. Perplexity Pro: $20 / mo (and it can't see your inbox).

Every product listed above exists because the underlying data was hard to get. Newsletters give you that data for free. Inbin structures it.

Also worth building