WeChat Chat Export Analyzer — Group Insights & Action Items

Analyze user-provided WeChat chat and group export records for recurring questions, needs, concerns, decisions, action items, supplied brand and competitor mentions, participation patterns, time patterns, and auditable conversation windows.

This Actor analyzes records the customer supplies. It does not log in to WeChat, scrape private chats, extract local databases, bypass encryption, control an account, or obtain chat records from a phone or computer.

Unofficial tool. This Actor is not affiliated with, sponsored by, or endorsed by Tencent, WeChat, or Weixin.

What the customer buys

One successful report costs $4.99 and analyzes up to 2,000 normalized messages.

The delivery contains:

• one privacy-aware row per normalized message
• transparent sentiment and intent labels
• exact supplied keyword, brand, product and competitor counts
• questions, needs and concerns with evidence
• decision and action-item signals with evidence
• nearby conversation windows around selected decisions and tasks
• participant, chat, daily and hourly activity patterns
• data-quality and coverage warnings

No usable messages means no report event is created.

This is not a chat extraction tool

The title deliberately says Chat Export Analyzer. Customers must first export, transform or otherwise lawfully obtain records they are authorized to analyze. Outside demo mode they must confirm hasRightToAnalyze.

Do not use this Actor to process private or confidential conversations without the required ownership, consent, authority and lawful basis.

Input modes

1. Static example preview

The default demo analyzes 12 clearly labeled synthetic messages across a project group and a customer-feedback group. It demonstrates the complete delivery without triggering the $4.99 custom report event. Apify may still show the small Actor-start or platform usage charge.

2. Existing Apify Datasets

{
  "mode": "datasetMessages",
  "sourceDatasetIds": ["YOUR_DATASET_ID"],
  "hasRightToAnalyze": true,
  "researchTopic": "Product launch customer feedback",
  "messageKeywords": ["价格", "退款", "发货", "副作用"],
  "knownEntities": ["产品甲"],
  "competitorNames": ["产品乙"],
  "privacyMode": "anonymizeParticipants",
  "redactContactDetails": true,
  "maxMessages": 2000
}

3. Uploaded JSON records

{
  "mode": "uploadedMessages",
  "hasRightToAnalyze": true,
  "researchTopic": "Campaign operations group",
  "uploadedMessages": [
    {
      "message_id": "m1",
      "chat_name": "Campaign group",
      "sender": "Member A",
      "timestamp": "2026-07-20 09:10:00",
      "content": "客户一直问价格和退款条件，怎么回答？"
    },
    {
      "message_id": "m2",
      "chat_name": "Campaign group",
      "sender": "Project owner",
      "timestamp": "2026-07-20 09:15:00",
      "content": "决定先更新FAQ，请内容负责人周五前提交。"
    }
  ]
}

Recognized fields

The normalizer recognizes common aliases:

• text: content, text, message, msg, body, message_text
• message ID: message_id, messageId, msg_id, msgId, localId, id
• sender: sender, sender_name, from_name, author, user, nickname, talker
• chat: chat_name, group_name, conversation, room_name, chat
• time: timestamp, time, create_time, created_at, date
• type: message_type, messageType, msg_type, type

Strings are also accepted as messages. Rows previously generated by this Actor are ignored when a Dataset is reused.

Privacy controls

Default settings:

• replace participant names with stable within-run labels such as Participant 1
• redact common email addresses, phone numbers and explicitly labeled WeChat IDs
• exclude common system, notification and media-placeholder rows

Customers may retain supplied names only when authorized. Redaction is a best-effort pattern screen, not a complete privacy, legal or security review. Review the delivered evidence before redistributing it.

Analysis behavior

The Actor uses deterministic lexical cues and exact customer-supplied terms. It does not claim to use AI and does not invent missing brands, people, decisions or tasks.

Message labels include:

• question
• purchase or budget signal
• recommendation request
• comparison or alternative
• complaint or service issue
• concern or objection
• decision
• action item or follow-up
• positive feedback or agreement

Discussion dimensions include price, quality, safety, trust, usage, delivery, support, alternatives, responsibility and time.

Important limits

• Findings describe only the records supplied to the run.
• Lexical labels can miss implied meaning, irony, slang, voice messages, stickers, images and context outside the supplied export.
• Action-item detection does not prove who owns a task or whether it was completed.
• Participant activity is not an employee-performance rating.
• Entity reporting covers only exact names supplied in the input.
• Time analysis depends on parseable timestamps and does not infer missing time zones.
• This Actor is not legal, compliance, medical, employment or privacy advice.

Billing

Main event:

chat-export-insights-report — $4.99

It is triggered once only when the complete report row is written. Row-level message classifications are then delivered without an additional Dataset-item event.