First request in five minutes

Quickstart

1. Get a token

   Generate one in the console under API after signup. The dbw_live_ prefix draws on your quota; dbw_test_ only reaches sample data and costs nothing.

2. Pick a source

   Browse the catalog, or call GET /v1/sources for the full list with health status.

3. Make the first call

   Three ways to do the same thing. All copy-paste.

curl -s "https://api.databridge.world/v1/feeds/tech-github-releases/items?limit=20" \
  -H "Authorization: Bearer $DBW_TOKEN"

import os, requests

r = requests.get(
    "https://api.databridge.world/v1/feeds/tech-github-releases/items",
    params={"limit": 20},
    headers={"Authorization": f"Bearer {os.environ['DBW_TOKEN']}"},
    timeout=15,
)
r.raise_for_status()
for item in r.json()["items"]:
    print(item["published_at"], item["title"])

const res = await fetch(
  "https://api.databridge.world/v1/feeds/tech-github-releases/items?limit=20",
  { headers: { Authorization: `Bearer ${process.env.DBW_TOKEN}` } }
);
const { items } = await res.json();
console.log(items.map((i) => i.title));

Auth

Every request carries a Bearer token in Authorization. Tokens are bound to your account and stop working the moment you revoke them.

Authorization: Bearer dbw_live_9f21c8d4e7a04b31

Endpoints

Response shape

Field names are identical across domains. Below is the full structure of one item.

{
  "feed": {
    "id": "tech-github-releases",
    "name": "GitHub · repository releases",
    "category": "tech",
    "updated_at": "2026-11-04T09:12:40Z"
  },
  "items": [
    {
      "id": "tech-github-releases:9f21c8d4",
      "title": "v2.4.0 — streaming parser rewrite",
      "url": "https://example.com/releases/v2.4.0",
      "summary": "Replaces the buffered reader with an incremental...",
      "content": null,
      "author": "example",
      "published_at": "2026-11-04T08:41:00Z",
      "updated_at": null,
      "lang": "en",
      "tags": ["release"],
      "hash": "9f21c8d4",
      "revision": 0,
      "deleted": false
    }
  ],
  "next_cursor": "eyJvIjoyMH0"
}

Paging and incremental fetch

Pull everything once, then only the delta. cursor is position-based rather than offset-based, so page 200 is as fast as page 1.

# Delta: only items after Nov 1
GET /v1/feeds/fin-cn-macro/items?since=2026-11-01T00:00:00Z

# Next page: pass the previous next_cursor back unchanged
GET /v1/feeds/fin-cn-macro/items?cursor=eyJvIjoyMH0

# next_cursor: null means you've reached the end

On dedupe: upsert keyed on id. Within a source it never changes; identical content across sources isn't merged for you, so compare hash yourself.

Quotas and rate limits

Quotas reset daily at 00:00 UTC+8. Live counters ride in response headers, so you can branch in code instead of guessing.

X-Quota-Limit: 20000
X-Quota-Used: 3412
X-Quota-Reset: 2026-11-05T00:00:00+08:00
X-Rate-Limit: 20          # requests per second
X-Rate-Remaining: 17

Exceeding quota returns 402; exceeding rate returns 429 with Retry-After. Neither costs money or suspends the account. Back off and retry.

Error codes

Webhooks (Premium)

Skip polling. New items POST to your endpoint on ingest, signed so you can verify authenticity.

POST https://your-app.example.com/hooks/dbw
X-DBW-Signature: sha256=6b1f...e0
X-DBW-Feed: tech-github-releases
X-DBW-Delivery: 01JC2F8Q

{ "feed_id": "tech-github-releases", "items": [ { "id": "...", "title": "..." } ] }

Signature = HMAC-SHA256(raw body, webhook_secret). Five retries with exponential backoff (10s → 160s); after that the delivery is logged in the console and can be replayed manually.