# Catch up on a busy Slack channel in one daily summary

Source: https://clickruntime.com/use-cases/summarise-slack-channel

> Pick the channel and when you want the summary. Each day ClickRuntime reads the new messages, has AI pull out decisions, questions and action items, and posts the summary where you'll see it.

## Can ChatGPT or Claude summarise a Slack channel for me every day?

They can summarise messages you paste in, but they won't read your channel every evening by themselves. With ClickRuntime connected, your assistant sets up an automation that reads the day's messages with your Slack connection, summarises them with the built-in AI, and posts the summary to a channel or emails it to you.

### Why the assistant can't do this on its own

- Reading a channel needs a Slack connection with read access that keeps working after the chat.
- A daily summary needs a schedule and a record of where the last summary stopped, so nothing is summarised twice.
- Summarising long threads is what AI is good at; the fetching, cutting and posting is plain code.

## Set it up from your AI assistant

Prompt to copy:

> Create a ClickRuntime automation: every weekday at 17:30 America/Chicago, read today's messages and thread replies in #product, and post to #product-digest a summary with three parts: decisions made, open questions, and action items with who owns them.

1. **Connect ClickRuntime to your assistant.** Add ClickRuntime as a connector (MCP server) in ChatGPT, Claude, Cursor or another MCP client, using https://clickruntime.com/api/mcp. You sign in with Google or email; there is no API key to copy. This is done once per assistant.
2. **Paste the prompt.** Start a new chat and paste: “Create a ClickRuntime automation: every weekday at 17:30 America/Chicago, read today's messages and thread replies in #product, and post to #product-digest a summary with three parts: decisions made, open questions, and action items with who owns them.” Change the details (times, channels, addresses) to yours.
3. **Read the draft.** The assistant restates what will happen, asks for anything missing, such as your timezone or which channel to post to, and saves a draft automation. A draft never runs on its own, so nothing happens until you agree.
4. **Connect accounts and save keys.** Reading needs a Slack account connected in the ClickRuntime dashboard (the workspace's own Slack app) with permission to read channel history; the assistant gives you the button. Posting the summary uses the notifications Slack.
5. **Test it.** The assistant runs it once for today so you see the summary, then tunes the format with you.
6. **Turn it on.** The assistant tells you what will run, when, and which services it will contact, then asks your permission to turn it on. From then on it runs on ClickRuntime's servers, and you can close the chat.
7. **Check it in the ClickRuntime dashboard.** Each run shows how many messages were read and that the summary was posted. State holds the timestamp of the last message summarised.

Connector setup for each assistant: https://clickruntime.com/assistants

## What the assistant writes

`slack_channel_digest.py`

```python
import time
from slack_sdk import WebClient
from clickruntime import on_trigger, credentials, state, notify, ai, log

SOURCE = "C0123PRODUCT"        # #product channel id
TARGET = "#product-digest"

@on_trigger
def run(event):
    client = WebClient(token=credentials.connection("slack", "work"))
    oldest = state.get("last_ts") or str(time.time() - 86400)
    msgs = []
    for page in client.conversations_history(channel=SOURCE, oldest=oldest, limit=200):
        msgs.extend(page["messages"])
    msgs = [m for m in msgs if m.get("type") == "message" and not m.get("bot_id")]
    if not msgs:
        notify.slack("No messages in #product today.", channel=TARGET)
        return {"messages": 0}

    users = {}
    def name(uid):
        if uid not in users:
            users[uid] = client.users_info(user=uid)["user"]["real_name"]
        return users[uid]

    lines = []
    for m in sorted(msgs, key=lambda m: float(m["ts"])):
        lines.append(f"{name(m.get('user', ''))}: {m.get('text', '')[:500]}")
        if m.get("reply_count"):
            replies = client.conversations_replies(channel=SOURCE, ts=m["ts"], limit=50)["messages"][1:]
            lines += [f"  ↳ {name(r.get('user', ''))}: {r.get('text', '')[:300]}" for r in replies]

    transcript = "\n".join(lines)[:60000]
    try:
        summary = ai.generate(transcript,
                              system="Summarise this Slack channel's day in three short sections: Decisions, Open questions, "
                                     "Action items (with owner). Skip small talk. Use Slack formatting.",
                              max_output_tokens=800).text
    except ai.AIError as e:
        log.warning("AI unavailable", {"code": getattr(e, "code", None)})
        summary = f"{len(msgs)} messages today in #product. AI summary unavailable."
    notify.slack(summary, channel=TARGET)
    state.set("last_ts", max(m["ts"] for m in msgs))
    return {"messages": len(msgs)}
```

- Reading uses your connected Slack account through slack_sdk; posting uses notify.slack.
- Bot messages are skipped and long messages are cut before they reach the AI.
- The last message's timestamp is kept in state, so tomorrow's summary starts where today's stopped.

## A schedule, not a trigger link

A daily schedule at the end of the working day, or first thing in the morning for the day before.

## Limits

- Private channels are readable only if the connected Slack account is a member.
- Each AI call takes about 32,000 input tokens; very busy channels are trimmed to fit, or summarised in two parts.
- Only the channels you name are read.

## Troubleshooting

### It fails with not_in_channel or missing_scope.

The connected Slack account can't read that channel. Join the channel, or reconnect Slack with history access in the ClickRuntime dashboard.

### The summary misses thread replies.

Ask the assistant to include threads; it fetches replies for messages that have them.

## Related

- [Slack automations](https://clickruntime.com/apps/slack)
- [Notion automations](https://clickruntime.com/apps/notion)
- [Linear automations](https://clickruntime.com/apps/linear)
- [AI triage and summaries](https://clickruntime.com/use-cases/ai-classify-and-summarise)
- [Daily Slack message](https://clickruntime.com/use-cases/daily-slack-message)
- [Daily standup in Slack](https://clickruntime.com/use-cases/daily-standup-in-slack)
- [Assistant scheduled tasks alternative](https://clickruntime.com/alternatives/chatgpt-claude-scheduled-tasks)
- [Zapier alternative](https://clickruntime.com/alternatives/zapier)
- [Docs: connections-and-secrets](https://clickruntime.com/docs/connections-and-secrets)
- [Docs: state](https://clickruntime.com/docs/state)

## Extend it: what you can add next

The first version is rarely the last. Because the automation is a Python script with memory, AI and any API available, you grow it by asking your assistant for the change. Each change is a new version, and a failed deploy never replaces the working one.

- **Summarise several channels in one digest, one section each** (Multi-step logic)
  Prompt: "Change my Slack digest so it also covers #design and #support, with one section per channel."
- **Create Linear issues for each action item** (Any app or API)
  Prompt: "Add to my Slack digest: create a Linear issue for each action item, assigned to its owner when the name matches."
- **Archive each day's summary to a Notion page** (Any app or API)
  Prompt: "Add to my Slack digest: add each day's summary to my Notion database Channel notes."
- **Flag messages that mention you and still have no answer** (Built-in AI)
  Prompt: "Change my Slack digest so it lists questions addressed to me that nobody answered."
- **A weekly version for people who only check on Fridays** (Schedules)
  Prompt: "Add to my Slack digest: every Friday at 16:00, post a summary of the whole week to #product-digest."
- **Track how many decisions and open questions there are each week** (Memory between runs)
  Prompt: "Add to my Slack digest: count decisions and open questions each day, and show the week's totals on Fridays."

## Questions

### Can ChatGPT summarise Slack channels automatically?

With ClickRuntime connected, yes. ChatGPT sets up an automation that reads the channel each day and posts an AI summary.

### Can Claude read my Slack and summarise it daily?

Yes, through ClickRuntime, using a Slack account you connect in the ClickRuntime dashboard.

### Does it read all my channels?

No. Only the channels named in the automation, and only those the connected account can see.

### Do I need an AI API key?

No. The summary uses ClickRuntime's built-in AI, including on the free plan.

### What can I add later?

More channels, Linear issues for action items, a Notion archive or a weekly version. See Extend it.
