> For the complete documentation index, see [llms.txt](https://docs.omnichat.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.omnichat.ai/en/release-note/sep-23-2026.md).

# Sep 23, 2026

Hello, dear Omnichat users!

🚀 Here are the latest  feature updates:

1. [New in OmniClaw : Read and Analyze Conversation Content](#new-in-omniclaw-read-and-analyze-conversation-content)
2. [Omni AI BYOA (Bring Your Own AI) Module Is Now Live](#omni-ai-byoa-bring-your-own-ai-module-is-now-live)
3. [WhatsApp Message Now Supports Free Entry Point (FEP) Windows](#whatsapp-message-now-supports-free-entry-point-fep-windows)
4. [WhatsApp Flow Creation Has Been Streamlined](#whatsapp-flow-creation-has-been-streamlined)
5. [Open API Pagination Upgrade (‼️ Action Required)](#open-api-pagination-upgrade-action-required)
6. [Insider Integration – Tag Sync Mechanism Updated](#insider-integration-tag-sync-mechanism-updated)
7. [New LINE Rich Menu URL Open Option](#new-line-rich-menu-url-open-option)
8. [WhatsApp Message Cost Estimates Now Use Your Billing Currency](#whatsapp-message-cost-estimates-now-use-your-billing-currency)

### New in OmniClaw : Read and Analyze Conversation Content

OmniClaw can now **read and analyze conversations directly within the platform** — surfacing message topics and customer intent, and letting you filter conversations by criteria to dig into root causes and get improvement suggestions.

Before analysis, conversations are automatically de-identified — **the AI only sees the content, not who said it, so customer personal data is never actually exposed to the AI model.**

**What you can use it for**

* **See what customers are asking about**: Want to know what's been coming up lately? Pick a date range and OmniClaw will break it down by topic, with the volume or share for each.
* **Find your best leads**: Looking for customers who seem ready to buy? OmniClaw picks out customers showing buying intent, tells you why it thinks so, and shows their last interaction time — so you can follow up right away.
* **Spot complaint patterns**: Want to know where complaints are piling up? OmniClaw finds conversations with complaints or negative sentiment, sorts and counts them, and shows which products, stores, or time periods they're linked to. You can even ask it for suggestions on a specific issue.
* **Filter and dig into causes**: Filter conversations by things like "response time too slow," a specific store, or a specific team — OmniClaw gives you the matching list and numbers, then helps explain what's driving it and how to fix it.
* **Set it and forget it**: Turn any of these into a Scheduled Task so OmniClaw runs them automatically — no need to ask every time.

<figure><img src="https://2848678129-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3u2I1osIy6mEi07X1z46%2Fuploads%2FfTkRLTMkBcoeVMacSc4m%2FScreenshot%202026-09-30%20at%202.37.59%E2%80%AFPM.png?alt=media&amp;token=81249b7c-6a34-40ae-858e-08bd72031cc5" alt=""><figcaption></figcaption></figure>

{% hint style="info" %}

* You can only analyze conversations you already have access to.
* Analysis only **covers the last 2 years of conversations** — OmniClaw will tell you directly if something's outside that range.
  {% endhint %}

### **Omni AI BYOA (Bring Your Own AI) Module Is Now Live**

Some businesses build their own AI agent in-house — often for security reasons or to handle specific use cases. That gives you full control over how the AI behaves, but usually means integrating it with every channel yourself: WhatsApp, your website widget, Messenger, and so on.

Now you can connect your own AI Agent straight to Omnichat. **Through the BYOA module's Open API, your AI Agent can plug into Omnichat's messaging platform** — once connected, it can receive customer messages, send messages, hand a conversation over to a human agent, or close a conversation.&#x20;

**Your AI handles the thinking; Omnichat handles connecting it to every channel** — so you don't have to build a separate integration for each one.

The BYOA module includes four APIs:

* `agent-messages` – send a 1:1 message
* `rooms/to-human-agent` – hand a conversation over to a human agent
* `rooms/close` – close a conversation
* `rooms/trigger-auto-assign` – send a conversation to a specific agent

Want to add BYOA? Just reach out to us!

<figure><img src="https://2848678129-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3u2I1osIy6mEi07X1z46%2Fuploads%2FWnY1KI8LkvpFm3gzxegL%2FScreenshot%202026-09-30%20at%202.38.39%E2%80%AFPM.png?alt=media&amp;token=6e77f3dd-b809-4ecc-a914-e841c1fc1f1f" alt=""><figcaption></figcaption></figure>

{% hint style="warning" %}
Each team can only add one Bring Your Own AI.
{% endhint %}

### WhatsApp Message Now Supports Free Entry Point (FEP) Windows&#x20;

When a customer messages you through a WhatsApp Click-to-WhatsApp Ad (CTWA) or a Facebook Page call-to-action button, and you reply within 24 hours, this opens a 72-hour "Free Entry Point" window — **during this window, you can send the customer any type of message for free.**&#x20;

With this update, **the conversation list and chat window will now clearly show whether a conversation is currently in this free window**, so you always know when you can message for free.

**How the 24-hour service window works**

* What it means: this applies when **a customer reaches out on their own (not through an ad) — it opens a 24-hour window for you to reply.**
* How it's counted: 24 hours starting from the customer's most recent message.

<figure><img src="https://2848678129-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3u2I1osIy6mEi07X1z46%2Fuploads%2FgTbisiZBxJJqsupyTCKc%2Fimage.png?alt=media&amp;token=b4ae5238-5a20-4cf6-b75e-b88733296626" alt=""><figcaption></figcaption></figure>

### WhatsApp Flow Creation Has Been Streamlined

Before, once you built a Flow in Meta's Flow Builder, you had to manually clean up the JSON before it would work in Omnichat — just copying and pasting it often meant only the first answer got saved. Now, you can paste a Flow built in Meta Flow Builder straight in and it just works — no manual JSON editing needed, and multi-step Flow forms are now supported too.

{% hint style="warning" %}

* Omnichat saves answers using each question's "label," so every label in the same form needs to be worded differently, or answers can get mixed up.
* Customer answers are saved by the Flow's "id," not its "title" — double check what each id actually refers to.
  {% endhint %}

### Open API pagination upgrade (‼️ action required)

This is a technical change to the Open API that needs action from your side. If your team uses our Open API, please pass this along to your developers. (We emailed affected customers about this on 2026/09/21 — please check your team got it.)

These endpoints now have a new v2 version with a better way of loading pages of data (cursor-based instead of offset-based):

* **`GET v1/messages`** (get message history)
  * v2 available: 2026/09/23 · v1 turns off: 2027/01/20
* **`GET v1/broadcast`** (get broadcast list)
  * v2 available: 2026/09/23 · v1 turns off: 2027/01/20
  * Note: for WhatsApp, `channelId` in v2 now returns the WABA number (v1 returned an Omnichat-generated ID instead)
* **`GET v1/broadcast/{id}/recipients`** (get broadcast recipient list)
  * v2 available: 2026/09/23 · v1 turns off: 2027/01/20
* **`GET v1/contacts`** (get channel contacts)
  * v2 available: 2026/10/21 · v1 turns off: 2027/01/20
  * v2 splits this into `GET v2/contacts` (list contacts) and `GET v2/contacts/lookup` (look up one contact)
  * Now also returns `fbBusinessId` (Meta Business Manager ID)
  * `subscribedAt` has been renamed to `createdAt`
* **`GET v1/tagging-logs/records`** (get tagging logs)
  * v2 available: 2026/11/04 · v1 turns off: 2027/01/20

### Insider Integration – Tag Sync Mechanism Updated

Before, when Omnichat synced tags to Insider, tags only ever got added — even if a tag was removed or changed in Omnichat, Insider would still keep the old one. Now, each sync fully replaces the tags with whatever's current, so Insider always matches what's actually true in Omnichat.

**🔍 Before VS. After**

**Before (tags just pile up):**

* 1st sync: `tag a`, `tag b`, `tag c` → Insider shows: `tag a`, `tag b`, `tag c`
* 2nd sync: `tag b`, `tag c`, `tag d` → Insider shows: `tag a`, `tag b`, `tag c`, `tag d` (old tags never removed)

**🆕 After (tags get replaced):**

* 1st sync: `tag a`, `tag b`, `tag c`→ Insider shows: `tag a`, `tag b`, `tag c`
* 2nd sync: `tag b`, `tag c`, `tag d` → Insider shows: `tag b`, `tag c`, `tag d` (matches the latest sync exactly)

### New LINE Rich Menu URL Open Option

Before, tapping a URL button on a LINE rich menu always opened it in an external browser, taking users out of LINE and making it harder to get back to the chat. Now, you can choose to have it **"Open within LINE"** instead.

<figure><img src="https://2848678129-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2F3u2I1osIy6mEi07X1z46%2Fuploads%2FcPCWArJoquIBD7G1KFfb%2FScreenshot%202026-09-30%20at%204.29.20%E2%80%AFPM.png?alt=media&amp;token=1788d0df-6ebe-44c8-b38e-33c6ed4f9f5f" alt=""><figcaption></figcaption></figure>

### WhatsApp Message Cost Estimates Now Use Your Billing Currency

The "estimated message cost" in the broadcast report used to only work in HKD, TWD, or USD, using fixed rates which made it hard to match up against your actual bill.&#x20;

Now, cost estimates use your team's own billing currency **(HKD, MYR, SGD, TWD, USD supported)** and the same WhatsApp rates used for your monthly bill, so it's much easier to check your numbers.

{% hint style="warning" %}
This only applies to broadcasts sent after this update — it won't be applied to past data.
{% endhint %}