Feishu Bitable Shortcuts for RedNote, Douyin and 1688 Data

To get Xiaohongshu (RedNote), Douyin, JD, Taobao or 1688 data into Feishu Base (also called Bitable), paste the links into one column and add an EveryInfra field shortcut next to it. You enter your EveryInfra API key once. The shortcut then fills every row with likes, saves, transcripts, prices or minimum order quantities. You need no scraper and no platform login. Failed calls and empty results cost nothing.

On October 22, 2026 we ran the first 5 shortcuts in a real Feishu Base table. The table values matched direct API calls. We ran the other 5 shortcuts against the live API on the same day, and all 5 returned data. This guide lists what each shortcut reads, what it writes back, what one row costs, and what each cell message means.

Listing status (at the time of writing): All 10 shortcuts are published in the Feishu Base developer center with "only visible to me" scope. Public listing requires Feishu developer onboarding, which is not complete yet. Once it is complete and the shortcuts go public, you can find them by name in the Shortcut Center. The shortcut names on the listing page are final.

Feishu or Lark? The shortcuts are published in Feishu, the China edition of the suite. Lark, the international edition, uses the same Shortcut Center design (Lark Help Center). If your team works in Lark, get the same data through the API or MCP setup at the end of this article.

What a field shortcut is

A field shortcut packs an AI, formula or API task into a field type. You configure it once, and it processes the whole column (Feishu official article, in Chinese). The EveryInfra shortcuts are API shortcuts. For each row, the shortcut reads one link, calls one EveryInfra capability, and writes structured fields back into that row. We built them on the official Feishu field shortcut template. Authentication is a Bearer key that each user enters in the field settings.

In one sentence: an EveryInfra field shortcut is a Feishu Base column that takes a platform link, returns that platform's public data, and charges your EveryInfra balance per row.

The 10 shortcuts: input, output, price

We read the prices from the live capability catalog on 2026-10-22 at 02:18 (Singapore time) and found them unchanged on 2026-10-24 at 18:33. The Taobao row has since been updated to the price in force from 2026-10-28 ($0.0045 per call; it was $0.001389). The pricing page is the reference. Each shortcut also adds a "Status" field: on success it shows the charge, on failure it shows the reason. The listing marks the shortcuts as free. The only charge is the per-row charge on your EveryInfra balance. Feishu's built-in AI shortcuts use a separate monthly AI quota (Feishu Help Center, in Chinese).

Three limits to know first:

• Douyin video data has no view count, because Douyin does not show views publicly (field notes in the Douyin scraper API docs). Bilibili does show views.
• JD, Taobao and 1688 prices are the price at read time. The shortcuts do not build price history.
• The RedNote comments shortcut reads only the first page and joins it into one cell. For all comments on a note, page through the API (batch comment guide).

Test 1: five columns in a real table

The test table is our own "EveryInfra field shortcut test" base. We ran it around 00:00 on 2026-10-22 (UTC+8). You must paste the key once per shortcut, because authorization does not carry over between shortcuts.

Test 2: the five new shortcuts against the live API

The new 5 shortcuts share code with the first 5. We called the live API once for each with a test key and formatted the result in the shortcut's columns. Times are 00:54 (first three) and 02:03–02:04 (Taobao, 1688) on Oct 22, Singapore time.

Test 3: empty results, offline cases and total spend

1. Counts are public counts at read time. The RedNote note showed 3,993 saves and 104 comments in the table. A direct API call a few minutes later returned 3,994 and 105. The Douyin video had 99,673 likes on Oct 21 and 99,689 on Oct 22 at 02:20. Record the read time when you compare snapshots (snapshot method).
2. Empty results are free. We also entered Bilibili BV1GJ411x7h7. The API returned an empty result and refunded the reserve. The cell said "no data found … not charged", and the charge was $0 (refund reconciliation).
3. 78 of 78 offline cases passed. Before release we ran 78 offline cases for the 10 shortcuts with the official template's local debug mode. The cases cover success, share text, empty rows, unknown links, missing key, invalid key, low balance, rejected parameters, empty results, paused capabilities and rate limits. The real-table run cost $0.014169 in total (including one Taobao detail call at that day's price of $0.001389).

Set it up in 4 steps

1. Create an API key. Sign up at everyinfra.com and create a key in the console (quickstart). Calls are pay per use. When the balance is low, the cell tells you.
2. Open the Shortcut Center. In grid view, click "+" right of the last field. In the field panel, hover over "Explore field shortcuts" and click "Field Shortcut Center" (same path as the Lark guide linked above). After public listing, search "Xiaohongshu", "Douyin", "JD", "Taobao" or "1688" and click "Use". Until onboarding is complete and the shortcuts are public, search does not find them.
3. Configure. Select the column that holds the links and paste your API key in the authorization field. The key stays in the field settings. It does not go into any cell.
4. Test a few rows, then the full column. Check a few rows first. The result is an object field, so you can expand likes, saves and other sub-fields into their own columns, then sort, filter and chart them. Use a new column: generating a full column overwrites the existing content of that column, and you cannot undo it (Lark and Feishu help centers both warn about this).

Two input notes. Keep xsec_token and the other parameters in RedNote links, and prefer links you copied from search results recently. JD short links (3.cn, u.jd.com) are not accepted: open them in a browser first, then copy the product page link. You need edit permission on the table (manage permission if advanced permissions are on).

Cell messages and charges

Error codes and retry advice: API error handling. Live availability per capability: status page.

Cost example: a 700-row research table

At the Oct 24 catalog prices (the same as on Oct 22):

• 200 RedNote notes: 200 × $0.005556 = $1.11
• 100 RedNote creator profiles: 100 × $0.005556 = $0.56
• 100 Douyin creator profiles: 100 × $0.000556 = $0.06
• 300 Douyin videos (data only): 300 × $0.000556 = $0.17
• Total: about $1.89 for 700 rows.

Rows that fail or return no data are free, so the real number can only be lower.

Other ways to get the same data

The shortcut is one row in, one row out. "Search a keyword and return 50 notes" is a one-to-many task, so a shortcut cannot do it. Use the RedNote search API or an AI client for that. Shortcuts also do not re-run on a schedule.

Same data in code or in an AI client

All 10 shortcuts call one endpoint. To read one 1688 product from a terminal:

curl -X POST https://api.everyinfra.com/api/v1/social \
  -H "Authorization: Bearer $EVERYINFRA_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"platform":"alibaba1688","action":"product_detail","params":{"url":"https://detail.1688.com/offer/674035283676.html"}}'

Response fields behind the 1688 columns:

In Claude, Cursor or another client that supports remote MCP, add this block and the AI can call the same capabilities with the tool everyinfra_call_api:

{
  "mcpServers": {
    "everyinfra": {
      "url": "https://api.everyinfra.com/mcp",
      "headers": { "Authorization": "Bearer <API Key>" }
    }
  }
}

Setup steps: remote MCP server. RedNote only: RedNote MCP.

Recent changes worth knowing (September–October 2026)

• Sep 15: Feishu 8.0. Agents can join group chats and work on Base tables. The Feishu CLI grew from 247 to 767 functions since March, and the call success rate rose from 78% to 95% (InfoQ, in Chinese). For table users this means an agent reads your table directly, so clean number and date columns matter.
• Oct 20: QuestMobile Q3 AI app report. AI office tools reached 142 million active users in September, up 50.6% year on year (36Kr summary). Tables are now a normal input for office agents.
• Oct 12, 20:00: JD Singles' Day sales opened (Ebrun, in Chinese). Prices move often during the event. Re-run the JD and Taobao price columns daily into new columns to keep your own snapshots.
• Sep 15: Cloudflare split crawler controls into Search, Training and Agent (Cloudflare blog). Sites now treat "an agent fetching a page for a person" differently from training crawlers. Read public data through a service that respects platform limits, and keep only the fields you need.
• Sep 7: China's Supreme People's Court issued its opinion on AI disputes (court announcement, in Chinese). It says that using AI to track and analyze a specific person's public information in order to pry into privacy infringes privacy rights. For market research, keep business fields and do not build files on individuals.
• Sep 9: the same court published 9 unfair-competition cases. One case found that mass-producing AI-written reviews of a rival's software to pull traffic is unfair competition (Xinhua, in Chinese). Keep this in mind if you turn collected data into comparison content.

FAQ

Do I need a RedNote or Douyin account, or cookies? No. The authorization field takes only your EveryInfra API key.

**Why should RedNote links keep xsec_token?** The API notes recommend full links copied recently from search results, with xsec_token kept. Links without it may return no data. No data means no charge.

Can a shortcut collect 50 notes for one keyword? No. A shortcut maps one link to one row. Use the search API or an AI client for keyword searches, then write the results into your table.

Related reading

• RedNote comments in batches
• RedNote analysis: do not look only at totals
• Remote MCP server