Bring WeChat Official Account CRM and website chat together. Start with our shared WeChat channel without certifying your own account, then use AI Custom Studio to build forms, charts and tools around your team.
OperationalNeeds attention
2 conversations waiting
Prioritized by wait time and customer context
AI reply suggestion
Generated from approved knowledge and active policies
This batch is covered by a temporary refund policy. Verify the order date and status before explaining the refund path.
Explore the console with read-only synthetic examples. Demo content is currently in Chinese. No messages, model calls or charges.
WeChat Official Account CRM
Bring WeChat and website conversations into one workspace. Manage customer details, internal notes, ownership and replies with clear business boundaries.
No own-account certification cost
Apply for the ReplyNest shared channel. Once enabled, place your merchant QR code on your website; visitors follow our official account and talk to your team.
Your brand, your channel
Already have an eligible official account? Configure the message URL, Token and EncodingAESKey, then connect your account credentials through the secure setup.
What “zero cost” means
The shared channel avoids the need to certify your own official account; it does not mean all ReplyNest subscriptions or services are free. Shared access requires activation and is intended for initial use, with shared quotas and availability. Delivery remains subject to WeChat permissions, messaging windows and rate limits. Keep Web Chat as a fallback.
Read the setup guideWebsite support · Web Chat
Add a chat bubble to your website so visitors can ask without switching to WeChat. Your team handles the conversation in ReplyNest, with enterprise knowledge ready to help.
Can I use this without a WeChat account?
Yes. Create a website channel, then install the embed code on your site.
Illustrative human reply · not a live conversation
Create a website channel, copy the embed code and ask your website maintainer to install it. No WeChat account is required.
Website and WeChat conversations enter the same workbench, with customer context, assignment and team-only private notes.
Start with human replies or reviewed AI drafts. Administrators can separately authorize knowledge-based automatic reception for the website channel.
Currently one website channel per tenant, counted toward the plan channel limit. Requires an HTTPS website you can edit. Installing the widget does not enable automatic AI replies; AI usage consumes points under the current plan rules. Browser storage affects visitor continuity, and a visitor name is not verified identity.
Business knowledge · RAG
Upload product information, service policies and FAQs to your business knowledge base.
RAG retrieves relevant knowledge first, then uses it to help draft an answer, with references for review.
Use assisted replies or drafts, or authorize low-risk knowledge answers. Insufficient evidence or high-risk questions go to an agent.
Knowledge and access are scoped to each business. Retrieval helps improve relevance; it does not guarantee that every answer is correct.
AI notes · Customer insights
Turn authorized conversation context into team-only notes and useful aggregates, not assumptions about sensitive identities.
Incrementally organize stated needs across conversations, alongside a short summary of the current issue.
Search AI notes in the support workspace. Agents can correct or delete AI results without overwriting manual notes.
Home dashboard charts summarize recent topics, stated needs and service preferences from saved results.
Disabled by default. Administrators separately authorize this purpose and set a daily AI-points budget. Generation consumes AI points; viewing saved charts does not call the model again or charge additional points. Multi-label needs and preferences use shares of category assignments, not exclusive shares of customers.
AI Custom Studio
Describe what your team needs. Build forms, record tables, charts and conditional workflows from supported capabilities, then review and publish a version you control.
Collect after-sales requests, consultation notes and internal tasks using fields you choose.
Group your application's records into bar or date-based line charts. Count, total or average numeric fields.
Set submitted fields with conditional rules, then add published forms to the support workspace.
An example you can ask AI to build
Create an after-sales form with a product, amount and priority. Mark amounts of 500 or more as priority, and chart totals by priority.
After-sales form
Totals by priority
Normal
Priority
Illustrative configuration · not live customer data
Review before publishing. Data stays scoped to the business and application. AI generates supported configuration, not arbitrary code; unsupported actions are not enabled by describing them.
Beyond a shared inbox
Connect channels, evidence, decisions, replies and alerts around the work support teams actually do.
Bring WeChat Official Account and website-chat conversations into one support workspace, with channel identity and routing kept traceable.
Authorize low-risk answers grounded in published knowledge. Questions with insufficient confidence or higher risk go to an agent.
Keep durable knowledge separate from temporary policies. Expired rules leave the decision context automatically.
Use browser push, team messaging tools and custom webhooks so urgent or unanswered conversations reach the right people.
Separate administrators, supervisors and agents. Member changes and sensitive actions remain auditable.
Tenant context follows retrieval, AI, notifications and audit. Sensitive integration secrets are never returned to the browser.
Three steps
Authorize a supported business account and verify event delivery.
Upload approved material and define temporary policy scope.
Agents review grounded suggestions, edit when needed and send.
RAG finds relevant knowledge. The Policy Engine checks effective dates and scope. Authorized history can become a knowledge draft, which requires administrator review and publication before use. Suggestions retain citations and feedback for review.
