Telegram New Posts Monitor — Recurring Public Channel Feed

Collect newly published posts from public Telegram channels into an Apify dataset, ready to export to JSON, CSV or Excel. The Actor saves a checkpoint between runs of the same task, so you can collect future posts without repeatedly exporting the same history.

中文用户 / 中文用戶： 简体中文说明 · 繁體中文說明（Telegram频道新帖监控、电报频道定时采集、Telegram频道发现）

Need a one-time channel export or accessible history? Start with Telegram Channel Scraper and its 10-post CSV/Excel starter. It exports one public channel as a spreadsheet snapshot, independent of this Monitor's checkpoint. Ten returned posts cost $0.002; from 19 November 2026, $0.00205 including the native start fee at the saved memory. The starter has a $0.01 run cap. Check its active Pricing tab before running.

Add a schedule and your own integration to send new posts to a spreadsheet, database or notification service.

No Telegram login, phone number, bot token or Telegram API credentials required. Collects the content publicly visible on each channel.

Ready-to-run public task

Open an example below and save your own task in Console. Adjust the visible inputs, check the active Pricing tab and maximum cost per run, then run your copy. Export the dataset as JSON, CSV or Excel after checking the run summary. Reuse that same saved task for later checks; a new task has an independent checkpoint.

Public task	Use case	Expected output
Monitor new public Telegram posts — Starter	Copy a small public-channel feed, then schedule the same saved task	Up to five recent posts on a new checkpoint, then only newer posts; a run with no new posts returns zero rows

Browse all ready-to-run examples → — ready-made tasks you can copy.

The example keeps reset disabled. Copy it as a new task for an independent checkpoint; SUMMARY reports each run's checked channels and errors. Actor and platform fees follow the active Pricing tab.

Monitor Telegram posts in three steps

1. Add Public channels and choose the New post limit per channel.
2. Keep Return recent posts, then new posts for an initial sample.
3. Save as a new task, run it, and schedule that same task for later collection.

The Console example requests up to five recent posts:

{
  "channels": ["cointelegraph"],
  "initialRun": "emitRecent",
  "maxPostsPerChannel": 5,
  "resetState": false
}

emitRecent selects the newest visible posts, up to the per-channel limit, and returns that initial sample oldest first. It applies only when that channel has no saved checkpoint in the current monitoring state. If it was already monitored, the run returns newer posts instead and can be empty. To get an independent first test, save a new task and run that task.

The first run inspects the latest public page as a starting sample. For one-time extraction or history, use Telegram Channel Scraper below.

From a Telegram history export to a recurring feed

Channel Scraper exports a snapshot; Monitor keeps its own checkpoint.

1. Choose the same public channels in both tools. For a handoff with overlap, first save a new Monitor starter task, use initialRun: "baseline", keep resetState: false, and run it. Confirm the channels were checked without errors.
2. Export the required history with Channel Scraper. SUMMARY and CHANNEL_REPORT show the covered ranges.
3. Run that same Monitor task again, then schedule it at a frequency suited to your channels. Run it one run at a time and review each run summary.
4. Merge both datasets using channelUsername + messageId as the unique key. The overlap is intentional. Confirm your destination saved each dataset independently.

Create the Monitor baseline before your export so the two datasets overlap. The new-post feed adds new posts; exported posts keep the text, views and reactions recorded at export time.

Set up a recurring feed

1. Save as a new task with your selected public channels.
2. Set initialRun to baseline to start silently from the current position, or emitRecent to include an initial recent sample. Leave resetState off.
3. Run the saved task once and inspect SUMMARY in its key-value store.
4. Schedule that same task at an interval appropriate for the channels’ posting frequency.
5. Export each run’s dataset or let your integration process it after completion.

A baseline run normally returns zero rows, and later runs return zero rows when there are no new visible posts. SUMMARY.errors lists any channel or discovery errors for the run.

Example input for a new task that starts with future posts only:

{
  "channels": ["cointelegraph", "the_block_crypto"],
  "initialRun": "baseline",
  "maxPostsPerChannel": 20,
  "maxPagesPerChannel": 5,
  "discoveryKeywords": [],
  "resetState": false
}

When a source error or an insufficient result budget leaves every checkpoint unchanged, the Actor skips rewriting the durable state. SUMMARY.checkpointPersistence reports unchanged or written; each run still exports its private MONITOR_STATE copy and summary. Successful idle checks continue to update the checked timestamp. An explicit reset is persisted even if the public checks fail.

Recover larger backlogs safely

Pending acknowledgements now use compact, exact ranges, so exceeding 1,000 saved pending posts no longer discards older acknowledgements. Gaps remain unacknowledged until those posts are actually saved. Existing checkpoints are read automatically on the next run; no reset is required.

Pending state holds up to 1,000 separate ranges per channel. When a scan would exceed that, SUMMARY.errors reports PENDING_CHECKPOINT_CAPACITY, and that channel keeps its previous checkpoint without saving additional posts. Increase the scan or post limits within their supported bounds to finish the backlog, or contact support with the run ID; each repeated run incurs its start fee. Keep the state rather than resetting it to clear this warning.

The compact checkpoint format requires build 0.3.14 or later. Do not switch a task with compact pending state back to an older build; use a compatible fix or complete the backlog on a supported build first. Keep your destination deduplicating by channel and message ID.

How checkpoints are separated

Each saved task has its own persistent monitoring state. Reuse the same task to continue its feed; a new task creates an independent starting point. Runs started directly from the Actor without a task share the account’s default monitoring state.

Changing initialRun does not replay a channel that already has a checkpoint. An empty input uses the small channel example and also keeps reset disabled. resetState: true starts fresh for the entire current monitoring state. Use it only when intentionally restarting: with emitRecent, previously returned posts can be returned again. Never leave reset enabled in a recurring schedule.

Store results in your destination using channelUsername + messageId as the unique key. Checkpointing avoids repeated collection; deduplicate retries and replays in your integration and confirm that it saved each dataset. Run one execution of the same task at a time, so each run starts from the previous checkpoint.

Save confirmation and retries

If saving a channel’s posts cannot be confirmed, its previous checkpoint is retained and the run fails after preserving SUMMARY. Look for POST_STORAGE_FAILED in SUMMARY.errors. Start a new run of the same task to retry. A partial or uncertain write can leave some posts in the failed run’s dataset; deduplicate failed-run and retry results by channel and message ID. Other channels that were successfully processed can retain their updated checkpoints.

When a partial scan has already saved posts, completing the backlog keeps the highest acknowledged post ID even if that post is temporarily not shown on the next check. This prevents that recovery case from replaying the post when it reappears. Keep deduplicating in your destination for resets and retries.

The checkpoint tracks collection, not delivery: confirm that your destination saved each dataset.

Handle busy channels and missed cycles

Input	Range	Purpose
maxPostsPerChannel	1–100	Maximum posts returned per channel in one cycle
maxPagesPerChannel	1–20	Maximum public pages inspected per channel with an existing checkpoint

When a cycle has more new posts than its limits, SUMMARY.channelsWithPendingBacklog lists the channels concerned. The previous committed checkpoint is retained while returned IDs are tracked as pending.

If the inspected pages cover the previous checkpoint but there are too many new posts for one cycle, later runs can return the remaining posts. If the page limit cannot reach the previous checkpoint, repeating the same settings may keep the backlog pending. Increase maxPagesPerChannel and consider increasing the post limit or running the task more often.

The scan window is set by maxPagesPerChannel. For a long interruption or a larger historical recovery, use a dedicated channel-history export and reconcile it with your destination before restarting the monitor’s baseline.

What each row contains

One canonical public Telegram post per row, with available fields such as:

• Source channel, message ID, public post URL and timestamp.
• Text, links, hashtags, mentions and forwarded-post label.
• Displayed views, structured reactions and media references.

Fields are null or empty when the channel shows no value publicly. Media are returned as references. Counts reflect the public display, including its rounding. The feed detects newer message IDs; each post keeps the text, views and reactions recorded when it was collected.

Discover public Telegram channels by topic

Add discoveryKeywords to find candidate public channels by topic:

{
  "channels": ["cointelegraph"],
  "discoveryKeywords": ["crypto news"],
  "maxDiscoveredChannels": 5,
  "initialRun": "baseline",
  "maxPostsPerChannel": 20,
  "maxPagesPerChannel": 5,
  "resetState": false
}

Discovery finds public channels related to the entered topics. It repeats for the entered topics on each run and adds platform usage costs. Candidates can change between runs: review useful candidates and add them to channels for stable monitoring.

countryCode and languageCode influence discovery preferences; they are not filters. Keep discovery empty for the simplest and most predictable workflow based on known channels.

Telegram monitoring API, n8n and Make workflows

Use case	Starting point	Destination
News and market announcements	A curated list of public news channels	Analyst review queue or database
Brand and competitor monitoring	Known company or industry channels	Filter returned text in your workflow
Research and AI indexing	Selected public sources	Deduplicated documents with source URLs
Source discovery	A small topic list plus known channels	Review candidates before adding them to monitoring

Use dataset exports or the Apify API with your HTTP client, Make, n8n or another integration. A finished-run webhook can trigger processing, including your own Telegram or email alerts. The task’s schedule starts collection, while your integration handles delivery and durable storage of the results.

Pricing

The active plan, effective October 17, 2026 at 11:19:23 UTC, charges $1.80 per 1,000 new posts saved, plus $0.00005 per Actor start event (one per GB of memory, minimum one) and separate platform usage.

At 128 MB, five saved new posts cost $0.00905 in Actor events, plus platform usage. An empty baseline or a cycle with no new posts has no post fee, but still incurs the applicable start fee and platform usage. Discovery can add usage costs; keep it empty when channels are already known. These are event-price examples; platform usage is billed separately. Check the Pricing tab for the active rate.

Set Maximum cost per run, a timeout and appropriate collection limits before scheduling. When result billing is active, the Actor checks the remaining result allowance before saving an entire channel batch. If the allowance is too small for that batch, collection stops, SUMMARY.chargeLimitReached is true, and SUMMARY.budgetStop reports the channel, required posts and remaining allowance. That channel's previous checkpoint is retained.

You can lower maxPostsPerChannel or increase the allowed event budget in the next run. The event budget covers Actor events; platform usage is billed separately and controlled by collection limits and the timeout.

If fewer result charges are confirmed than the requested batch, the run ends as failed without advancing the affected checkpoint; deduplicate its output on retry. An empty baseline is expected behavior.

Chinese Simplified

中文说明：Telegram新帖监控工具（Telegram New Posts Monitor 爬虫 / API）

Telegram New Posts Monitor 定时采集公开Telegram频道（电报频道）新发布的帖子，保存到Apify数据集，可导出为JSON、CSV或Excel。同一个任务在多次运行之间自动保存检查点（checkpoint），之后每次只采集新帖子，不必反复导出相同的历史。还可以按主题发现新的公开频道。无需Telegram登录、手机号、Bot Token或Telegram API凭据。

适合需要持续跟踪新闻频道、行业频道和竞品频道的分析师、品牌舆情监控团队，以及需要稳定数据源的AI和数据工程团队。

为什么选择这款Telegram频道监控工具？

• 自动保存检查点：每个保存的任务都有独立的监控状态。重复运行同一个任务，就能继续它的新帖流；新建任务则从新的起点开始。
• 两种首次运行方式：Get recent posts, then watch for new ones 先返回最近的帖子样本（按从旧到新排列），之后只返回新帖；Start watching now — no historical posts 只记录当前位置，从未来的帖子开始。
• 积压帖子分批追赶：频道一次发帖过多时，未完成的部分会记录为待处理，后续运行继续返回；SUMMARY.channelsWithPendingBacklog 列出相关频道。
• 预算保护：保存一个频道的批次之前先检查剩余费用额度；额度不足时停止采集，并保留该频道原有的检查点。
• 按主题发现频道：填写 discoveryKeywords，查找相关的公开频道候选，审核后再加入固定监控列表。
• 接入你的工作流：通过Apify API、运行完成的Webhook以及n8n、Make、Zapier或HTTP工作流，把新帖子送到表格、数据库或通知服务。

适用场景

• 新闻和市场公告：监控一组精选的公开新闻频道，把新帖送入分析师审核队列或数据库。
• 品牌和竞品舆情监控：跟踪公司或行业频道，在你的工作流中按正文筛选新帖。
• 研究和AI索引：从选定的公开来源获取去重后的帖子，并保留原帖链接。
• 频道发现：用少量主题词加上已知频道，找到候选频道后再加入监控。
• 从历史导出到持续采集：先用 Telegram Channel Scraper 导出历史，再用同一组频道启动本工具，按 channelUsername + messageId 合并两个数据集。

可以获取哪些数据？

每行是一条规范的公开Telegram帖子，包含可用的字段：

• 来源频道、消息ID、公开帖子链接和发布时间。
• 正文、链接、话题标签、提及和转发来源标签。
• 显示的浏览量、按表情统计的表情回应和媒体引用。

每次运行的键值存储中还有 SUMMARY，汇总已检查的频道、错误、待处理积压和检查点写入情况。

价格

自2026年10月17日11:19:23（UTC）起生效的计费方式：

计费项目	价格
每1,000条保存的新帖子	1.80美元
每次Actor启动	0.00005美元（每GB内存一次，至少一次）
平台用量	另行计费

在128 MB内存下，保存5条新帖子的Actor事件费用为 0.00905美元，另加平台用量。基线运行或没有新帖的运行不收帖子费用，但仍有启动费和平台用量。启用频道发现会增加平台用量；已经知道频道时可以留空。可在 Run options 中设置 Maximum cost per run、超时和采集上限。请以当前Pricing页面显示的价格为准。

使用方法

1. 在 Channels to monitor 中添加公开频道，每行一个：cointelegraph、@cointelegraph 或 https://t.me/cointelegraph。
2. 设置 New post limit per channel，建议先用5条。
3. 在 First run 中选择先返回最近帖子，或只从未来的帖子开始。
4. 保存为新任务（Save as a new task），运行一次并查看 SUMMARY。
5. 为同一个任务设置定时运行（Schedule），频率按频道发帖速度设定；每次运行后导出数据集，或交给你的集成自动处理。

控制台示例（最多5条最近帖子，之后只采集新帖）：

{
  "channels": ["cointelegraph"],
  "initialRun": "emitRecent",
  "maxPostsPerChannel": 5,
  "resetState": false
}

只从未来帖子开始的示例：

{
  "channels": ["cointelegraph", "the_block_crypto"],
  "initialRun": "baseline",
  "maxPostsPerChannel": 20,
  "maxPagesPerChannel": 5,
  "discoveryKeywords": [],
  "resetState": false
}

也可以直接打开现成的 新帖监控示例任务，复制为自己的任务后再设置定时运行。

常见问题

会发送Telegram提醒吗？ 本工具负责采集数据集。把它连接到你自己的n8n、Make、Zapier或HTTP工作流，即可推送提醒。

为什么第一次运行没有帖子？ 选择 Start watching now — no historical posts 时，首次运行只建立基线，通常返回0条。已经监控过的频道会沿用原有检查点，只返回更新的帖子。SUMMARY.errors 和预算设置可以说明其他空结果的原因。

多个任务的检查点会互相影响吗？ 不会。每个保存的任务都有独立的监控状态；不通过任务、直接运行Actor时，使用账号的默认监控状态。

需要Telegram机器人或API密钥吗？ 不需要。只需公开频道的用户名或链接。

定时运行时要注意什么？ 保持 resetState 关闭，重复运行同一个任务，不要让同一任务的运行重叠；在目标系统中按 channelUsername + messageId 去重。

中文说明会翻译采集到的数据吗？ 不会。中文说明只是使用文档；采集到的数据保持来源的原始语言（帖子正文按频道发布时的原文呈现）。

其他Telegram工具

• Telegram频道采集工具（历史导出CSV）：一次性导出公开频道的帖子和可访问的历史，每1,000条0.20美元，支持CSV、Excel、JSON和Markdown。
• Telegram关键词搜索工具：不知道该看哪些频道时，按关键词、品牌名或产品名查找Telegram公开帖子，适合品牌提及和舆情监控。

Chinese Traditional

中文說明（繁體）：Telegram新貼文監測工具（Telegram New Posts Monitor 爬蟲 / API）

Telegram New Posts Monitor 定期擷取公開Telegram頻道（電報頻道）新發布的貼文，儲存到Apify資料集，可匯出為JSON、CSV或Excel。同一個任務在多次執行之間自動儲存檢查點（checkpoint），之後每次只擷取新貼文；也可以依主題發現新的公開頻道。不需要Telegram登入或API憑證。

為什麼選擇這款Telegram頻道監測工具？

• 自動儲存檢查點：每個儲存的任務都有獨立的監測狀態。
• 兩種首次執行方式：Get recent posts, then watch for new ones 先回傳最近的貼文；Start watching now — no historical posts 只從未來的貼文開始。
• 依主題發現頻道：填寫 discoveryKeywords，審核後再加入監測清單。

適用場景

• 新聞追蹤：監測精選的公開新聞頻道。
• 品牌輿情監測：追蹤公司或產業頻道的新貼文。
• 競品監測：持續追蹤競品頻道。
• 社群研究：從選定的公開來源取得去重後的貼文。

使用方法

1. 在 Channels to monitor 中新增公開頻道，每行一個：cointelegraph、@cointelegraph 或 https://t.me/cointelegraph。
2. 設定 New post limit per channel（建議先用5則），並在 First run 選擇首次執行方式。
3. 儲存為新任務（Save as a new task），執行一次並查看 SUMMARY，再為同一個任務設定排程（Schedule）。

{
  "channels": ["cointelegraph"],
  "initialRun": "emitRecent",
  "maxPostsPerChannel": 5,
  "resetState": false
}

現成範例任務：新貼文監測範例。

價格

自2026年10月17日11:19:23（UTC）起：每1,000則儲存的新貼文 1.80美元，每次Actor啟動 0.00005美元（每GB記憶體一次，至少一次），平台用量另行計費。在128 MB記憶體下，儲存5則新貼文的Actor事件費用為 0.00905美元，另加平台用量。

常見問題

會傳送Telegram通知嗎？ 不會。連接你自己的n8n、Make、Zapier或HTTP工作流程即可推送通知。

為什麼第一次執行沒有貼文？ 選擇 Start watching now — no historical posts 時，首次執行只建立基線，通常回傳0則。

中文說明會翻譯擷取到的資料嗎？ 不會。說明是中文，資料保留來源的原始語言（貼文內容以頻道發布時的語言呈現）。

完整欄位與輸出範例請見英文說明或简体中文说明。

Which Telegram Actor should I use?

Your goal	Recommended Actor	Start with an example
Find public posts by keyword	Telegram Keyword Search — discovery with literal matches in text or channel title Product mentions → CSV
Export recent posts or accessible history	Telegram Channel Scraper — date filters and CSV, JSON or Markdown layouts Known channels → CSV
Collect newly published posts repeatedly	Telegram New Posts Monitor — reuse the same saved task and checkpoint New posts → saved checkpoint

All three work with public Telegram channels, with no Telegram login.

Frequently asked questions

Does this Actor send Telegram alerts?

It collects a dataset. Connect your own n8n, Make, Zapier or HTTP workflow to deliver alerts.

Why did the first run return no posts?

Future posts only creates an empty baseline. A channel already monitored in this task uses its existing checkpoint. SUMMARY.errors and the budget fields explain any other empty run.

Does it track edited posts or updated view counts?

It detects newer message IDs; each post keeps the text, views and reactions recorded when it was collected.

Do I need a Telegram bot or API key?

No. Enter public channel usernames or links.

Run quality and diagnostics

SUMMARY.quality reads bounded when the attempted public scan finished, limited when a selected limit, budget or collection scope bounded the run, and partial when a source error, a known coverage gap, a pending backlog or a row without a required post ID, link or timestamp occurred. recordsInspected, missingRequiredFields and optionalFieldsUnavailable are field checks, not billing counters.

A channel whose public content cannot be read is reported as CHANNEL_UNREADABLE and keeps its checkpoint, so no false empty baseline is created. Initial baselines and recent samples use the channel's current public posts, and channelsNotChecked lists unprocessed channels.

If older posts cannot be reached, the Actor stops that channel scan and reports PAGINATION_NOT_ADVANCING in SUMMARY.errors. Successfully saved new posts are kept as pending IDs; the previous committed checkpoint remains in place until a later scan catches up. Reuse the same saved task with reset disabled to resume without re-emitting those pending IDs. SUMMARY.pagination reports the scan (pagesRead steps read, nonAdvancingPages stopped scans); these are not billable events.

The Actor checks each parsed post's channel username before applying the post limit, saving rows or updating a checkpoint. Posts belonging to another channel are excluded, reported as UNEXPECTED_POST_CHANNEL and counted in SUMMARY.sourcePostsIgnored. If a channel shows only posts from other channels, the previous checkpoint is retained and no initial baseline is created. Forwarded text and links to other channels inside a valid post are kept in its available fields.

How is an empty channel discovery reported?

Only an explicit no-results answer is treated as a verified empty search. A discovery result that cannot be verified is reported in SUMMARY.errors as DISCOVERY_UNVERIFIED. A discovery-only run with no channels and a discovery error ends as failed; explicitly configured channels are still monitored and the run reports the discovery error. Duplicate topics differing only in letter case are searched once per run. For stable recurring feeds, save the useful channel usernames and leave topic discovery empty.

Real observed output excerpt

Reduced record from an owner verification run observed on 17 October 2026. Field values are retained from that observation and may have changed since then.

{
  "channelUsername": "cointelegraph",
  "messageId": "72692",
  "messageUrl": "https://t.me/cointelegraph/72692",
  "publishedAt": "2026-10-16T00:56:04.000Z",
  "views": "7.98K",
  "scrapedAt": "2026-10-17T17:33:17.174Z"
}

Post text, reactions and media are omitted here. Publication time precedes observation time: an initial emitRecent run returns the latest visible posts, which can be older. The formatted views string is preserved. Later cycles return only unseen posts; keep the same saved task to continue from its baseline.