StayPresent

🛖 About

A lightweight Python package designed to keep your bots and background scripts alive by running a dedicated Flask web server alongside your main application(s).

Perfect for deploying on platforms like Render, Railway, Koyeb, Heroku, or any host that requires an active HTTP port to keep your service running.

Contents

• Features
• Installation
• Quickstart
• Usage Guide
  • Text Response
  • JSON Response
  • HTML Response
  • Markdown Response
  • Custom Host, Port, and Threads
• Custom Paths & Multiple Responses
• Running Multiple Bots
• Self-Ping / Keep-Warm
• API Reference
• Logging
• Requirements
• Use Cases

📖 Full Documentation: For complete documentation, API reference, deployment guides, and FAQs, open DOCUMENTATION.md.

✨ Features

• Zero-Friction Setup: Get running with just one line of code.
• Production-Ready by Default: Automatically detects and uses waitress to avoid Flask's "development server" warnings.
• Multiple Bot Support: Run several bot processes side-by-side under one web server, each monitored and restarted independently, with shared or per-bot arguments/environment variables.
• Auto-Restarts & Crash Recovery: Automatically respawns any bot process if it crashes, complete with customizable delays and consecutive-crash budgets — per bot.
• Flexible Responses: Serve custom plain text, JSON (default), full HTML templates, or beautifully rendered Markdown.
• Built-in Markdown Renderer — Zero Extra Dependencies: Headings, emphasis, strikethrough, links, images, autolinks, nested lists, nested blockquotes, tables, fenced/indented code blocks, raw HTML passthrough, and a GitHub-flavored stylesheet — no markdown package required.
• Light / Dark / Auto Theming: Every Markdown page can be forced to light or dark, or left on auto (default) to automatically follow the visitor's own OS/browser preference.
• Page Metadata for Markdown Pages: Set a custom favicon, <title>, and meta/Open Graph description directly from Python — no template editing needed.
• Custom Paths, Multiple Responses: Host more than one response at once at different paths (e.g. /, /status, /dashboard) — handy for multi-bot setups where each bot gets its own endpoint.
• Static Asset Serving: Automatically serves CSS, JS, images, and favicons located next to your HTML/Markdown files, scoped correctly even when several are hosted at different paths.
• Advanced Process Control: Easily pass custom command-line arguments and environment variables directly to your bot process(es), shared across all of them or configured individually per bot.
• Fail-Safe Logging: Logs a clear, dedicated error if the underlying web server dies unexpectedly, and never touches Python's root logger.
• Optional Self-Ping / Keep-Warm: Periodically ping your own public URL in the background to prevent free-tier hosts from spinning your service down due to inactivity — fully opt-in, off by default.

📦 Installation

Standard Installation:

pip install staypresent

Production Installation (Recommended):

To suppress Flask's development-server warning and use a production-grade WSGI server, install the prod extra. This automatically provisions waitress.

pip install staypresent[prod]

(Note: If waitress isn't installed, StayPresent gracefully falls back to Flask's built-in development server and logs a one-time warning. Nothing else is required — Markdown rendering, tables, theming, and everything else described below works out of the box with no additional packages.)

🚀 Quickstart

import staypresent

staypresent.run("bot.py")

That's it. This starts a background web server (defaulting to 0.0.0.0:8080, serving {"message": "I'm Present"} at /) and runs bot.py alongside it, automatically restarting it if it ever crashes.

📘 Usage Guide

By default, if you don't configure anything via staypresent.web, the root path (/) returns a JSON response: {"message": "I'm Present"}. Every function below accepts an optional path argument to host more than one response at once — see Custom Paths & Multiple Responses.

Text Response

import staypresent

staypresent.web.text("Service is Online")
staypresent.run("bot.py")

JSON Response

import staypresent

staypresent.web.json({
    "status": "online",
    "uptime": "24h"
})
staypresent.run("bot.py")

HTML Response

Serve a full HTML file. Any CSS/JS/images referenced next to it are served automatically.

import staypresent

staypresent.web.html("template/index.html")
staypresent.run("bot.py")

Markdown Response

Serve a .md file, rendered to clean, styled HTML — headings, emphasis, links, images, lists, blockquotes, tables, and fenced code blocks are all supported out of the box, with no extra dependency required. The file is re-read (and re-rendered) fresh on every request, so editing it on disk shows up immediately.

import staypresent

staypresent.web.markdown("CHANGELOG.md")
staypresent.run("bot.py")

The rendered page uses a GitHub-flavored stylesheet and automatically matches the visitor's light/dark preference. You can also customize the theme and page metadata:

import staypresent

staypresent.web.markdown(
    "docs/guide.md",
    path="/docs",
    mode="dark",                              # "light", "dark", or "auto" (default)
    favicon="favicon.png",                    # a file next to guide.md, or a direct URL
    title="Project Docs",                     # page <title> + Open Graph title
    description="Everything you need to get started.",  # meta + Open Graph description
)
staypresent.run("bot.py")

Files (images, etc.) next to your .md file are served automatically, exactly the same as with html().

Custom Host, Port, and Threads (Complete Example)

import staypresent

staypresent.web.json({"status": "running"})

staypresent.run(
    "bot.py",
    host="0.0.0.0",
    port=5000,
    threads=8
)

🧭 Custom Paths & Multiple Responses

Every staypresent.web function (text, json, html, markdown) accepts an optional path argument. It defaults to "/", matching the original behavior — but you can host as many independent responses as you like, each at its own path, all served by the same web server.

import staypresent

staypresent.web.json({"status": "online"})                       # served at "/"
staypresent.web.text("bot #2 is alive", path="/bot2")             # served at "/bot2"
staypresent.web.html("dashboard.html", path="/dashboard")         # served at "/dashboard/"
staypresent.web.markdown("CHANGELOG.md", path="/changelog")       # served at "/changelog/"

staypresent.run("bot.py")

Note on trailing slashes: for html()/markdown() at any path other than "/", StayPresent automatically redirects /dashboard → /dashboard/. This isn't optional — it's what makes relative asset links inside your file (<link href="style.css">, <img src="images/logo.png">, a relative favicon) resolve correctly against that file's own directory instead of its parent. text()/json() responses don't need this since they have no static assets to resolve.

A handful of small helpers make working with multiple paths easier:

staypresent.web.paths()                 # -> ['/', '/bot2', '/changelog', '/dashboard']
staypresent.web.get_all()               # -> {'/': {...}, '/bot2': {...}, ...}
staypresent.web.get("/bot2")            # -> {'type': 'text', 'value': 'bot #2 is alive'}
staypresent.web.remove("/bot2")         # stop hosting a response, returns True/False

Note on /health: StayPresent has a built-in default at /health returning {"status": "ok"} (see Built-in Health Check below). It's a default, not a reservation — if you register your own response at /health via staypresent.web, yours is served instead.

🤖 Running Multiple Bots

staypresent.run() can launch and independently supervise more than one bot process at once, all behind the same web server.

Same args/env for every bot

Pass a list of file paths instead of a single one:

import staypresent

staypresent.run(["telegram_bot.py", "discord_bot.py"])

Each bot is monitored and restarted on its own — one crashing (and getting restarted, per max_restarts/restart_delay) has no effect on the others. bot_args and env, if provided, are applied identically to every bot in the list.

Per-bot args/env

For finer control, use the bots argument instead — a list of dicts, one per bot:

import staypresent

staypresent.web.json({"status": "online"})

staypresent.run(bots=[
    {"file": "telegram_bot.py", "args": ["--verbose"]},
    {"file": "discord_bot.py", "env": {"SHARD": "0"}},
    {"file": "worker.py"},
])

bots is mutually exclusive with bot_file/bot_args/env — pick whichever style fits: bot_file (+ optional shared bot_args/env) for the simple case, bots when each process needs its own arguments or environment.

How failures are handled with multiple bots

• Each bot has its own independent restart counter, so max_restarts is a per-bot budget.
• staypresent.run() waits for every bot to finish before returning or exiting — it doesn't stop supervising the others just because one of them gave up.
• If any bot ultimately fails to stay up (restarts exhausted, or restart_on_crash=False and it crashed), staypresent.run() exits the whole process with a non-zero exit code once every bot has finished, the same fail-loud behavior as the single-bot case.
• Ctrl+C / SIGTERM stops the web server and all bot processes cleanly.

📡 Self-Ping / Keep-Warm

Some free-tier hosts spin your service down after a period of inactivity, even if staypresent.run() has an open port. staypresent.ping()/staypresent.cron() let you periodically hit your own public URL to keep it warm — entirely optional and off by default.

One-off ping

import staypresent

result = staypresent.ping("https://my-bot.onrender.com")
print(result)
# {'url': 'https://my-bot.onrender.com/', 'ok': True, 'status_code': 200, 'elapsed': 0.42, 'error': None}

Recurring keep-warm pings

import staypresent

staypresent.web.json({"status": "online"})

handle = staypresent.cron("https://my-bot.onrender.com", interval=300)  # every 5 minutes

staypresent.run("bot.py")

cron() returns a CronHandle you can use to stop it later:

handle.stop()               # stop the background pinger
handle.is_running           # True/False

You can also pass on_success/on_failure callbacks to react to each ping's result (e.g. for your own logging/metrics), and pass host="127.0.0.1", port=8080 instead of a full URL if you just want to ping your own local server.

⚙️ API Reference

staypresent.run(...)

Launch your bot script(s) alongside the web server.

Note: port, threads, max_restarts, restart_delay, and restart_reset_after are validated up front — passing an invalid value (e.g. threads=0, a negative port) raises a ValueError immediately instead of failing silently or deep inside waitress. Likewise, every bot file (in bot_file or bots) is checked to exist before the server starts.

Crash Recovery Details

StayPresent automatically monitors every bot process. If one exits with a non-zero exit code, StayPresent restarts it based on your configuration:

• Clean Exits: An exit code of 0 is considered intentional and will not trigger a restart.
• Independent Supervision: With multiple bots, each one is monitored and restarted completely independently — one crashing (or exhausting its restarts) doesn't pause or stop the others.
• Manual Shutdowns: Stopping StayPresent via Ctrl+C (SIGINT) or SIGTERM shuts down the server and every bot process cleanly.
• Smart Counters: The max_restarts limit applies to consecutive crashes, per bot. If a bot runs successfully for the duration of restart_reset_after (default 60 seconds), its crash counter resets.
• Non-Zero Exit on Giving Up: staypresent.run() waits for every bot to finish. If any bot ultimately failed to stay up — restarts disabled and it crashed, or max_restarts was exhausted for it — staypresent.run() then exits the whole process with a non-zero exit code instead of returning normally. This lets a hosting platform's own restart-on-crash policy (Render, Railway, Docker, systemd, etc.) kick in as a last resort, instead of the process quietly exiting 0 as if nothing went wrong.

staypresent.web

Built-in Health Check

A default /health endpoint is available out of the box, returning {"status": "ok"}. This is incredibly useful for platform pingers and uptime monitors that require a dedicated health-check path separate from your regular response(s). It's a default, not a reservation — registering your own response at /health via staypresent.web overrides it.

staypresent.ping(...) / staypresent.cron(...)

🪵 Logging

StayPresent logs to its own "staypresent" logger (via a dedicated StreamHandler), never touching Python's root logger — so it won't clobber, duplicate, or reformat log output your bot script has already configured for its own loggers. Startup, restarts, crashes, and shutdowns are all logged with timestamps at the appropriate level (INFO/WARNING/ERROR).

🛠 Requirements

• Python 3.8+
• Flask
• waitress (optional, but highly recommended for production — pip install staypresent[prod])

Markdown rendering, theming, tables, and everything else in this README works with no additional dependencies — StayPresent ships its own built-in Markdown-to-HTML renderer.

💡 Use Cases

• Keeping a Discord/Telegram/Slack bot alive on a free-tier host that requires an open HTTP port.
• Running several bots (e.g. a Telegram bot and a Discord bot) from a single deployed service.
• Exposing a lightweight status page, uptime dashboard, or CHANGELOG.md/README.md viewer for a background worker.
• Giving a hosting platform's health-check probe something to hit while your real work happens in a separate process.