How to Use the Podhoc Telegram Bot: Generate Podcasts From Chat

Generate AI podcasts from inside Telegram

To use the Podhoc Telegram bot, find it in Telegram, send /login to link your Podhoc account, then send /create <url> to generate a podcast from any public link. Optional flags — --lang, --duration, --style, --voices — fine-tune the output. The bot also exposes /podcasts to list your library, /wishlist and /add to manage saved URLs, and /credits plus /subscription to check your account state.

The bot is the same Podhoc generation pipeline you know from the web app, accessible via Telegram on any device. It is the fastest path from “I should listen to this article on the way home” to “I have a podcast version of it on my phone”.

Why a chat bot for podcast generation?

A chat interface is the lowest-friction surface for a tool you only use occasionally. You do not need to remember a URL, switch tabs, log in to a dashboard, or open another app. You see something interesting, you paste it into Telegram, you get audio back.

For users who already have Telegram open all day, the bot collapses three minutes of UI navigation into one line of chat. For users who do not — the bot is the moment Podhoc moves from “useful tool I sometimes remember to use” to “background utility that lives where I already work”.

Step 1 — Find the bot in Telegram

Open Telegram on any device — mobile, desktop or Telegram Web. Search for the Podhoc bot using the search bar at the top of the chat list. Tap Start to begin the conversation. The bot replies with a welcome message and a short list of available commands.

If you cannot locate the bot, the support page carries the current handle. The bot username may evolve as the integration matures, so the support page is the canonical source.

The bot is a thin wrapper around the same backend the web app and the Podhoc API hit. Anything you do in chat is reflected on your Podhoc account in real time — and vice versa. Generate a podcast on the bot and it appears in your web library; add a URL to your wishlist on the website and it shows up in /wishlist on Telegram.

Step 2 — Link your account with /login

Send /login to the bot. It replies with a magic-link URL pointing at app.podhoc.com. Tap that link on the device where you are already signed in to Podhoc (or paste it into your browser and sign in there). The link expires once used.

Once linked, the bot can:

• Read your credit balance with /credits.
• Show your current plan with /subscription.
• List your generated podcasts with /podcasts.
• Manage your wishlist with /wishlist and /add.
• Generate new podcasts with /create.

To unlink, send /logout. The bot keeps no token after that — re-running /login re-establishes the connection.

If you do not yet have a Podhoc account, the welcome message links to podhoc.com where the Free plan lets you sign up and try the product without a card.

Step 3 — Create your first podcast with /create

The headline command is /create. It takes a URL and optional flags:

/create https://example.com/article
/create https://example.com/article --lang es
/create https://example.com/article --duration 15
/create https://example.com/article --style feynman_technique --voices 3
/create https://example.com/article --lang fr --duration 20 --style debate --voices 2

The defaults are sensible — --lang your account default, --duration 10, --style deep_dive, --voices 2 — so a bare /create <url> is a complete request.

Generation takes 2-5 minutes. The bot replies immediately with a podcast_id and a “generating” status. When the episode completes, the bot pushes a follow-up message with a play link and a download link. You can also call /detail <id> to inspect any single podcast or /download <id> to fetch the MP3.

The eight --style values map to the eight Podhoc audio styles:

• deep_dive — two voices explore the topic comprehensively.
• didactic — structured teacher-style delivery.
• feynman_technique — first-principles re-explanation, ideal for active learning.
• critique — methodological and evidentiary evaluation.
• debate — multiple voices argue different positions.
• simplified_explanation — compressed orientation summary.
• pedagogical_framework — scaffolded teaching with explicit learning checkpoints.
• alchemist_formula — guided step-by-step transformation, ideal for procedural how-tos.

Pick the style that matches the source. Research papers benefit from critique. Tutorials benefit from didactic. Long articles benefit from deep_dive. The same source generates noticeably different output across the eight styles — try two on the same URL and the difference is obvious.

Step 4 — Manage URLs with /wishlist and /add

Not every URL is worth listening to immediately. The wishlist is your “save for later” surface.

/add https://arxiv.org/abs/2501.12345
/add https://example.com/long-article
/wishlist

/add <url> saves the URL to your account-wide wishlist. The wishlist is shared between the bot and the web app — anything saved on your phone shows up in your browser library, and vice versa. /wishlist lists everything saved, with the option to convert any entry into a podcast.

This pattern is useful when you spot a long read during the day but only have time to listen in the evening — drop it into the wishlist with a quick /add, then convert it during your commute home.

Step 5 — Check credits and inspect your library

/credits      → "You have 2,847 credits remaining."
/subscription → "Pro plan, billed monthly. Renews 2026-06-15."
/podcasts     → Paginated list of your last N podcasts.

/credits returns your remaining balance. Generation consumes credits according to the standard Podhoc pricing formula — max(30, ceil(duration × 2.5)) plus optional bonuses. A 15-minute, 2-voice, single-source episode is roughly 45 credits; a 30-minute, 3-voice, multi-source episode comes in around 200.

/subscription shows your current tier (Free, Creator, Pro). Most bot commands require Pro. /credits, /subscription, /login, /logout, /start and /help work on any tier.

/podcasts lists your library with titles, durations and IDs. Use the IDs to call /detail <id> or /download <id> for a specific episode.

The full command reference

Common patterns

A few things teams and individuals use the bot for repeatedly:

• Commute reading queue. Throughout the day, /add long articles you spot. On the commute home, /wishlist and /create the ones you want to hear.
• Multilingual practice. Pass --lang fr (or any of 74 languages) to generate the same source in a language you are learning. Listen alongside the original.
• Group context syncing. Add the bot to a Telegram group with your team and use it to generate audio briefings on whatever URL someone shared in chat.
• One-line API substitute. If you do not need the full API yet but want a chat-driven generation flow, the bot is the lowest-overhead path. See API integration ideas for when to switch.
• Keeping up with your own field. Drop an arXiv URL or a competitor blog post in chat, get a 15-minute briefing in audio.

For more concrete patterns, see Daily-life applications for the Podhoc Telegram bot.

What the bot does not do (yet)

• Group-wide auto-conversion. The bot does not currently watch group channels and convert every shared URL — every generation is still triggered by an explicit /create.
• File uploads. Like the API, the bot accepts URLs only. PDF / DOCX uploads are a web app feature for now.
• Voice messages as input. The bot reads text commands; it does not transcribe voice notes into source material.
• Inline mode. The bot operates in a private chat or group context; it is not an inline-mode bot you can invoke from any chat.

These gaps are honest constraints, not bugs. The web app fills the file-upload gap; the API fills the programmatic-bulk gap; the bot fills the chat-driven, single-URL, low-friction gap.

Privacy and security

The bot’s security posture borrows the discipline of the rest of the Podhoc backend. Bot tokens and webhook secrets are stored in AWS Secrets Manager and never logged. Webhook requests are verified using a per-call secret token included in the X-Telegram-Bot-Api-Secret-Token header — only the genuine Telegram service can talk to our backend.

Database queries from the bot use parameterised queries; no raw SQL string interpolation. The Lambda runs inside the VPC private subnets — same security boundary as every other Podhoc backend component. Privacy follows the main Podhoc privacy policy.

Try it now

Open Telegram, search for the Podhoc bot, send /start, and /login. Within a minute you can run /create <url> on the article you have been meaning to read. The first episode lands in 2-5 minutes — and from then on, your “I should listen to this” thoughts have a one-line home.

Sign up for Podhoc Pro →