Fang Labs

Every customer answered at once, at any hour. What matters reaches your team first.

Fang watches every comment and chat on your Facebook page around the clock, answers customers at once in your admins' own voice, and sends complaints to your team's LINE group before they spread.

Nothing to install · Cancel and delete your data any time · We help set up your first page

Fang's queue (sample data): a complaint at the top, orders passed to sales, and a general question waiting for someone to take it.
Sample data

In daily use at OSUKA

OSUKA's customer-care team, a Thai power-tool brand (~470K followers on Facebook + TikTok, Nov 2026), runs on the system Fang grew out of, every day: sorting comments, alerting LINE to a crisis, and tracking rivals' promotions.

  • 20–30 h →
    14 min
    Median time to first reply, before and after the systemOSUKA CS team, 11–17 Oct 2026
  • 10 / 10Crisis comments caught in the test setTest set of 34 messages, all 34 passed
  • 1,507Rival promotions the system found in 3,970 rival postsCounted to early Oct 2026
A case page in Fang (sample data): a customer comment about a refund, the team's internal note, the reply Fang drafted for review before sending, the case owner and the case history
Sample data

That night, with Fang

An illustrative scenario

With Fang

  1. Sunday night, 10 January

  2. Joy comments under the promotion post: “Ordered 5 days ago, still nothing. Is this a scam?”

  3. Fang replies at once, in the same voice as the page's admins: “Sorry for the wait. Could you send your order number in chat? Our admin will check it first.”

  4. Fang alerts the admin team's LINE group with the customer's words and a link to the case.

  5. With the complaints category switched on, Fang only acknowledges. The case stays open for your team.

  6. The customer sends the order number in chat. The case goes to the top of the queue.

  7. Monday morning

  8. When an admin logs in, the order number is already waiting, so checking the parcel and updating the customer comes first.

If nobody sees it

  1. More negative comments pile up under the post.

  2. A screenshot is posted in a buyers' group.

  3. Monday morning

  4. By the time an admin sees it, it has already been shared.

Where the admin team works

Comment and chat queue

Every comment sorted, all handled in one place

  • When a customer says it broke, Fang files a complaint, never a sales lead.
  • Price questions go to sales.
  • Every case has a reply deadline that runs on your business hours.
  • Chats ready to buy get their own sales queue.
Fang's queue (sample data): a complaint at the top, orders passed to sales, and a general question waiting for someone to take it.
  1. Negative comments rise to the top
  2. Anything waiting past your set time turns red
  3. Reply right here in Fang
Fang's Messenger view (sample data): a queue of ready-to-buy chats beside an order card built from the chat.
Product, quantity, payment, phone and address, pulled from the chat into one card
  1. Product, quantity, payment, phone and address, pulled from the chat into one card
  2. Shows how long you can still reply inside Meta's 24-hour window
  3. Once someone takes a chat, only they reply. No crossed answers.
  • See who has a case open, so nobody answers twice
  • AI drafts, an admin presses send
  • Take the 5 oldest cases in one click
  • End a shift and hand cases back, none left hanging
  • Internal notes and @mentions
  • Spot-check the quality of closed cases

Try the demo (The demo runs on sample data)

Auto-reply

Instant replies,
in your admins' own voice

In Fang, replies are not canned messages.
Fang learns from the replies your admins have actually sent: their sign-offs, their tone, the way they apologise. Then it answers customers at once, around the clock.

The shop's voice · version 1 · approved

  • Read from 184 real admin replies
  • Ends with “kha” in 73% of messages
  • Uses an emoji in about 12% of messages
  • Medium length, 2–3 sentences
  • Usually opens with “sawasdee kha”

You choose which kinds of message Fang answers.

  • Starts in test mode: logged, not sent
  • Goes live once you approve the voice
  • You decide how sure Fang must be before it answers alone
  • One click stops every auto-reply

Reply logSample dataTap the price to see the shop fact it came from

  1. Comment

    Do you have a branch near Ram Inthra?

    Thanks for asking! Our admin will check and reply in chat shortly.

    Sent · handed to the teamNot in the shop's facts

  2. Messenger

    My face went red and burning, a rash all over.

    Not answered · sent to the team at onceAuto-reply stays off for crisis

  3. Comment

    How much is the vitamin C serum?

    In stock! The 30 ml vitamin C serum is a bottle. Message us for more details.

    SentVoice version 1

Shop facts · Products and pricesVitamin C serum 30 ml: 590 baht a bottle · in stock

Every price, promo and promise in a reply comes from your own shop facts.

LINE alerts

LINE only buzzes
when a person needs to act.

Everyday questions wait in the queue, not in LINE, so the team's group stays quiet until something really needs a person.

  1. New complaint · as it comes in
  2. Yesterday's summary · 08:00 every morning
  3. Case close to its reply deadline · in business hours

Sent from your own LINE Official Account

Beyond your Page

Brand listening

What customers say about your brand,
everywhere they say it.

Fang gathers the public reviews and comments about your brand on Shopee, Lazada, Pantip, TikTok, YouTube and in the news, and sorts them into topics: what customers praise, and what they complain about.

Every number says how many posts it counts, and opens to the posts behind it.

Refreshed every week, with a monthly summary · Public data only · Deleted on the writer's request

Customer voice · last 30 daysSample data

  • Shopee412
  • Lazada236
  • TikTok188
  • YouTube63
  • Pantip41
  • News9

TopicPraiseComplaints

  1. Shopee ★2

    Arrived 3 days late, and the box was dented

  2. Lazada ★5

    So fast. Ordered yesterday, here today

  1. Pantip

    Great value on promo, a bit steep at full price

  2. TikTok

    Cheaper on the live than in the shop

  1. Shopee ★5

    Three months in and it's as good as day one

  2. YouTube

    The new batch isn't the same, the texture changed

  1. Lazada ★3

    The lid didn't seal and it spilled in the box

  2. Shopee ★5

    Packed really well, double bubble wrap

  1. Pantip

    Is this shop genuine? It's much cheaper than elsewhere

  2. News

    The brand posts how to check a product is genuine

Counted from 949 posts · tap a topic for the posts behind it

Your own Facebook Page and Instagram account connect through Meta's official route

Competitors Sample data

Know what rivals' customers complain about before you plan products and content.

Fang gathers the comments under rivals' public posts and sorts the complaints by brand and by product, with the same rules it uses on your Page.

Fang's competitor voice page (sample data): complaints by product, ours against rivals'
Rivals' complaints by product. Products with too few comments say “not enough data”.
Fang's page comparison (sample data): our Page against the rivals' median
Your Page against the rivals' median. Pages that would skew it are left out of the median.
Fang's competitor content page (sample data): products rivals post more about, and their promo posts
Which kinds of rival post work, and which ones your Page hasn't tried yet.

Competitor data comes from public posts · Rivals' comments never enter your case queue

KOL Sample data

Same fee. Not the same views.

Paste the clip links and see cost per view by creator, by clip and by product, before you book the next round.

  • One board, 5 steps. Anything stuck for 7 days rises to the top.
  • Creators' phone, LINE, email and fee are visible to admins only
  • TikTok views refresh on a schedule; other platforms are entered by hand
Fang's creator table (sample data): three creators on the same fee, with different costs per view
  1. Invited3 stuck
  2. Agreed2 stuck
  3. Shipped1 stuck
  4. Posted1 stuck
  5. Paid

For owners

Executive overview · Per-person numbers · Chat quality Sample data

Know if the team answers on time, who is swamped, and which chats turned into sales.

Everyone can see who is handling what, so nothing is answered twice or missed. It covers comments and Messenger chats, and owners get a monthly report of sales that came from chat.

Fang's executive overview (sample data): what to look at first, and sales from cases this month
What to look at first, which team has it, and the sales your team logged from closed cases, against the same days last month.
Fang's per-person numbers (sample data): replies, reply times and closed cases for each person
Each person's replies and reply times. Use it to talk with the team, not to judge people.
Fang's chat quality (sample data): the score, its 5 parts and what to fix first
A quality score in 5 parts (every chat answered, answered fast, hand-overs picked up, customers keep talking, sales closed) and what to fix first · Messenger chats only

Flags a bot that still mentions an expired promo

Ask AI

Ask like you would ask your team.
Every number opens to its rows.

Every number is counted from your own comments, chats and cases. AI only picks the data and writes the opening line.

Negative comments per day · last 7 daysSample data

27negative in the period

▼ 1 against the 7 days before (28) · peak Thursday, 8

2Sat0Sun4Mon6Tue6Wed8Thu1Fri

Open the listcounted from 27 rows

Which case has waited longest right now?Sample data

CustomerWaiting

  1. Crisis Two days of the sunscreen and I've a red rash all over my face
  2. Ordering 4 serum sets, cash on delivery
  3. Do you have a branch near Ram Inthra?

Open the listcounted from 16 rows

6 ready-to-buy chats waiting: who's waited longest?Sample data

WantsWaiting

  1. Serum set ×3 · bank transfer
  2. Serum set ×1 · cash on delivery
  3. Sunscreen ×2 · bank transfer

Open the listcounted from 6 rows

Did we answer on time yesterday?Sample data

  • 36 of 42replied
  • 86%replied (%)
  • 25 minfirst reply (median)
  • 21slower than target

Open the listcounted from 42 rows

Where Fang comes from

Every page, on every platform, answering its customers in time.

20–30 hoursthe wait for a first reply, before Fang

14 minuteswith Fang

*Median time to first reply of the customer-care team at our pilot organisation, before and after Fang. Results vary with each business's message volume and how its users work.

[To be added: a real, approved quote from a pilot shop]

[To be added: who said it][To be added: their role], [To be added: shop name]

Fang started from a real problem inside an organisation: customers who got in touch on social media waited 20 to 30 hours for an answer, and many other business pages on Facebook have the same problem.

So we built a system that sorts every comment, puts what is urgent first and alerts the team's LINE group to complaints. Once it was in daily use, the median wait for a first reply fell to 14 minutes.

Our goal is the same as on day one: every customer should get an answer without waiting, whichever way they get in touch. Today Fang answers on Facebook and Messenger and listens on Shopee, Lazada, Pantip, TikTok, YouTube and the news, and more channels will follow.

[To be added: founder photo]

Varit Juramongkol (Nano)Founder, Fang Labs

Your customers' data stays yours. Check every point yourself.

  1. Connect with Facebook Login. No page password.

    Connected pages
  2. Connection data is stored encrypted, every business's data is kept apart, and there's an encrypted backup every night.

  3. Disconnect a page and Fang deletes its data at once · Set your own retention period

    Settings › Business
  4. Every reply, assignment, hide or delete, and settings change is logged · Hiding and deleting stay off until an admin turns them on · Admins and team members have separate rights

    Settings › History
  5. Fang strips out phone numbers, emails, addresses and customer names before a message goes to OpenAI's AI, and those messages are not used to train OpenAI's models.

  6. We'll sign a PDPA data processing agreement with you.

Connected pagesSample data

  1. Kleebbua SkincareFacebook Page · connected

    Deletes every comment, chat and history record of this page, at once.

  2. Kleebbua Skincare, SiamFacebook Page · connected

    Deletes every comment, chat and history record of this page, at once.

Read our privacy policy

Common questions

Who is Fang for?

Business pages with comments and chats every day that need to answer customers in time, from a shop with one admin to brands running several pages. The question categories adapt to what you sell.

Which channels does it cover?

Facebook Page comments and Messenger chats come into the queue and are answered in Fang. Brand listening gathers public reviews and comments from Shopee, Lazada, Pantip, TikTok, YouTube and the news.

How is it priced?

Plans start at 2 accounts with unlimited users, and you pick only the features you need. Add accounts or features any time. Ask our team for a quote.

What does Fang answer on its own?

Whatever your shop has entered: products and prices, FAQs, opening hours, shipping, returns and warranty policies, and the promotions that are live. For a parcel's status, Fang asks for the order number and hands the chat to your team. Customers get answers in their own language, and when Fang is not sure enough, the reply waits as a draft for your team.

Can I connect more than one page?

Connect as many pages as you run, all in one business workspace.

Where do the voices from other channels come from?

Fang reads public data only: product reviews, forum threads, comments under clips and news that anyone can open. It never uses anyone else's account and never reads private messages.

Does it still work with AI off, or out of AI credit?

The queue keeps running and Fang sorts with keyword rules instead. Auto-reply pauses and cases wait for your team.

How do admin and team-member rights differ?

Admins connect pages, change settings, switch auto-reply on or off and see the history. Team members work the queue, take cases and reply. Creators' contacts and fees are for admins only.

Does it stop if we go over our plan?

No. Fang warns you before you reach the limit, and our team will suggest a plan that fits better.

Plans and pricing

From 2 accounts, unlimited users, the features you choose.

Pay only for the features you use, such as auto-reply, Messenger chats, brand listening and KOL, and add pages or features any time.

Request a quote

Connect your page today. Tonight's comments won't wait for morning.

Up and running in 3 steps

  1. Connect your pageLog in with Facebook and pick the page.

  2. Connect your team's LINELink your shop's LINE Official Account to the admin team's group.

  3. Switch on auto-replyTest it first (logged, not sent), then approve your shop's voice.

Book a demo