Skip to main content

GoyGram

GoyGram Logo

Python 3.11+ Rust Core License: AGPL v3 Telegram API Security Docs & Wiki

What is this?

Ultimate split-brain Telegram framework (Python + Rust core) built for production-grade speed, control, and maximum OpSec.

Under the hood: a Python orchestration layer drives two completely independent network transports (Bot API over aiohttp + MTProto over raw TCP with full DH key exchange), both feeding into a single async event bus. Every crypto operation — AES-256-IGE for MTProto packets, AES-256-GCM for session vaults — runs in a Rust .so compiled with LTO and opt-level=3. Hand-written TL codec, no code generation at runtime. QR code login rendering in the terminal via qrcode + Rich. SRP password proofs for 2FA. And the vault: your auth key locked to your machine-id through PBKDF2-SHA256 at 600,000 iterations.

Key Features

  • Split-brain architecture: ergonomic Python layer + blazing-fast Rust extension.
  • Session Eater: aggressive in-memory cleanup (zeroize strategy for legacy .session files after migration).
  • Vault AES-256-GCM: encrypted local session bootstrap. Key derived from machine-id + session name via PBKDF2 (or bypass with GOYGRAM_VAULT_KEY).
  • TUI auth flow: terminal-first authorization workflow — phone login with SMS code, QR code scanning in ASCII art, 2FA/SRP password challenges. All Rich-styled when a TTY is present.
  • Proxy support: SOCKS5 (with user/pass auth) and HTTP CONNECT tunneling for MTProto connections. Also respects ALL_PROXY / HTTPS_PROXY / HTTP_PROXY env vars.
  • Dual transport: Bot API (HTTP long-polling via aiohttp, multipart uploads, auto-webhook-clear on 409) + MTProto (raw TCP with AES-256-IGE, dynamic salt recovery on bad_server_salt, auto-DC migration on PHONE_MIGRATE_N) — in one app runtime.
  • DC Routing: MTProto uses a built-in map of the five Telegram DC endpoints and selects the preferred DC, falling back to 149.154.167.50:443 (DC 2).
  • Dynamic API dispatch: every Bot API method works via __getattr__sendAnimation, getUserProfilePhotos, setMyCommands, whatever. Snake_case auto-converts to CamelCase. mt_ prefix routes to MTProto.
  • Keyboard system: inline keyboards, reply keyboards, force reply, reply removal. All with to_dict() serialization that adapts per transport.
  • Forum topic management: full create/edit/close/reopen/delete lifecycle for forum topics and the General topic. Both transports supported.
  • Zero-copy event objects: MsgObj, CbObj, PollObj, MemberObj with __slots__ — no per-message dict overhead.
  • Composable filters: boolean AND/OR/NOT on Filter (filters.text & ~filters.me).
  • Multi-session: named vaults (session_name="worker_1") for farming multiple accounts from the same process. Separate auth keys, separate TCP connections, separate self_id.

Installation

pip install goygram

Requires Python 3.11+. Pre-built wheels ship for Linux, Windows, macOS, and FreeBSD where the corresponding runner build succeeds. Termux is natively validated in a Termux environment; install the Python package from source there because Android/Termux wheels are not interchangeable with manylinux wheels. Rust is not required for the standard Linux, Windows, and macOS wheels. Installs aiohttp, rich, and qrcode as dependencies.

FreeBSD and Termux

FreeBSD packages are built by the release workflow inside a FreeBSD 15 VM and attached to the GitHub Release because PyPI rejects FreeBSD's nonstandard wheel platform tag. The Rust core is built in the official termux/termux-docker environment and attached as a native validation asset; Termux users should build locally from the source distribution. On a real Termux device, install the Termux toolchain and build from the source distribution:

pkg update
pkg install python rust clang
python -m pip install --no-build-isolation .

Quick Start

1) Bot API (token)

import asyncio
from goygram import GoyGram, filters

app = GoyGram(bot_token="123456:ABC_TOKEN")

@app.on_msg(filt=filters.text)
async def echo(msg):
    await msg.reply("Hello from Bot API")

asyncio.run(app.run())

2) MTProto (no bot token, requires API ID + API Hash)

import asyncio
from goygram import GoyGram

app = GoyGram(api_id=123456, api_hash="0123456789abcdef0123456789abcdef")  # auto-fetches Telegram DC endpoint at startup

@app.on_cmd("ping")
async def ping(msg):
    await msg.reply("pong from MTProto (api_id/api_hash)")

asyncio.run(app.run())

3) Named MTProto sessions (multi-session in one folder)

import asyncio
from goygram import GoyGram

app = GoyGram(
    api_id=123456,
    api_hash="0123456789abcdef0123456789abcdef",
    session_name="farm_worker_1",
)

asyncio.run(app.run())
  • By default, session data is stored in default.vault.
  • With session_name="farm_worker_1", session data is stored in farm_worker_1.vault.
  • If farm_worker_1.session exists, it is migrated to farm_worker_1.vault during bootstrap (securely zeroized after).

Dynamic API & Methods

GoyGram can route Bot API method names dynamically, including methods that are not hardcoded as convenience methods:

  • Call Bot API methods directly even if they are not explicitly hardcoded:
    • await app.sendDocument(chat_id=..., document=...)
    • await app.getChat(chat_id=...)
    • await app.getUpdates(timeout=30)
  • Snake-case also works and is converted to Bot API method names:
    • await app.send_document(chat_id=..., document=...) -> sendDocument
  • MTProto actions (authorized with API ID/API Hash) are available with mt_ prefix:
    • await app.mt_get_dialogs(limit=50)
    • await app.mt_get_chat_full(chat_id=...)

This behavior is implemented through dynamic method resolution in the client core (__getattr__) and transport-aware request routing.

Authentication & Security

Interactive Login

On first run with MTProto, GoyGram launches a Rich-powered TUI:

GoyGram Interactive Login

? Choose login method:
  > QR Code Login
    Phone Number Login

Choose QR code (scan with any Telegram client) or phone number (SMS code). 2FA password is handled automatically via SRP proofs. The resulting session is stored as default.vault — AES-256-GCM encrypted, keyed to your machine.

Vault Encryption

  • Algorithm: AES-256-GCM (authenticated encryption via Rust's aes-gcm crate)
  • Key derivation: PBKDF2-HMAC-SHA256, 600,000 iterations, key material = {machine-id}:{session_name}
  • Override: GOYGRAM_VAULT_KEY env var (base64-encoded 32 bytes) bypasses PBKDF2 entirely The vault does not fall back to silently accepting plaintext after a failed decryption.

Session Migration

Telethon/Pyrogram .session files are auto-detected, read from SQLite, migrated to .vault, and securely zeroized (overwrite + fsync + unlink).

Developer Tools (Help)

Use built-in introspection tools:

app.help()            # pretty DX overview in console
print(dir(app))       # inspect available shortcuts + dynamic entries

or:

from goygram.utils import print_methods
print_methods(app)

With type hints on key event objects (MsgObj, CbObj, MemberObj, PollObj) and filter primitives, modern IDE autocomplete works much better out of the box.

Filters

goygram.filters supports composable boolean operators:

from goygram import filters

smart_filter = filters.text & ~filters.me
another = filters.text | filters.me

@app.on_msg(filt=smart_filter)
async def handler(msg):
    await msg.reply("Filtered")

Built-in filters: filters.text (message has text), filters.me (message from current account/bot). Compose with &, |, ~. Custom filters: Filter(lambda e: ...).

Transport Routing

Messages can be routed explicitly by transport:

# Force Bot API
await app.send_msg("bot:123456789", "via bot", via="bot")

# Force MTProto
await app.send_msg("mt:123456789", "via mt", via="mt")

Chat ID prefixes (bot: / mt:) are auto-resolved. When replying, the transport source is preserved automatically — reply to a Bot API message, it goes back via Bot API.

FSM Persistence

The default FSM remains in memory:

app = GoyGram(bot_token="123456:ABC_TOKEN")

For an external store, pass an object with load() and save(snapshot) methods:

class RedisFSM:
    def __init__(self, redis):
        self.redis = redis

    def load(self):
        return self.redis.json().get("goygram:fsm") or []

    def save(self, snapshot):
        self.redis.json().set("goygram:fsm", ".", snapshot)

app = GoyGram(bot_token="123456:ABC_TOKEN", fsm_backend=RedisFSM(redis))

For complete control, use fsm_on_change. It receives a JSON-compatible snapshot after every state change and can write it to Redis, PostgreSQL, a file, or another service:

def persist_fsm(snapshot):
    external_store.write(snapshot)

app = GoyGram(bot_token="123456:ABC_TOKEN", fsm_on_change=persist_fsm)

The active core object is also available as app.fsm. It exposes snapshot() and restore(snapshot) for explicit checkpoints and migrations. Existing set_state, get_state, get_state_data, and clear_state behavior is unchanged.

Event Pipeline

BotNet.spin() ──→ bus.push("bot", data)
                                          ──→ Disp.consume() → your handlers
MTNet.spin() ──→ bus.push("mt", data)

Single asyncio.Queue → typed event objects (MsgObj/CbObj/PollObj/MemberObj) → handler lists in registration order. Per-handler error isolation — one crashing handler never takes down the dispatcher.

Logging

GOYGRAM_LOG=DEBUG python app.py   # verbose (raw MTProto packet dumps)
GOYGRAM_LOG=INFO python app.py    # default (startup, errors)
GOYGRAM_LOG=WARNING python app.py # quiet

Logger hierarchy: goygram.app, goygram.botapi, goygram.mtproto, goygram.disp, goygram.security, goygram.dc.

Architecture at a Glance

┌─────────────────────────────────────────────┐
│             GoyGram (Public API)             │  ← User-facing facade
├─────────────────────────────────────────────┤
│        AppCore (Internal Engine)             │  ← Config, hooks, routing
├──────────────────┬──────────────────────────┤
│ BotNet (aiohttp) │   MTNet (TCP/MTProto)    │  ← Independent transports
├──────────────────┴──────────────────────────┤
│          Bus → Disp (Event Pipeline)         │  ← asyncio.Queue + dispatcher
├─────────────────────────────────────────────┤
│  goygram.ext (Rust .so) — AES-IGE/AES-GCM   │  ← Native crypto (LTO, opt=3)
└─────────────────────────────────────────────┘

Wiki

📚 Official documentation and Wiki. There are separate pages for using the client, Bot API, MTProto, events, bytes and TL data. 👉 Open GoyGram Pages · Open GitHub Wiki

License

See LICENSE.

Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

goygram-0.7.47.tar.gz (78.1 kB view details)

Uploaded Source

Built Distributions

If you're not sure about the file name format, learn more about wheel file names.

goygram-0.7.47-cp311-abi3-win_amd64.whl (310.1 kB view details)

Uploaded CPython 3.11+Windows x86-64

goygram-0.7.47-cp311-abi3-manylinux_2_34_x86_64.whl (395.9 kB view details)

Uploaded CPython 3.11+manylinux: glibc 2.34+ x86-64

goygram-0.7.47-cp311-abi3-macosx_11_0_arm64.whl (369.0 kB view details)

Uploaded CPython 3.11+macOS 11.0+ ARM64

File details

Details for the file goygram-0.7.47.tar.gz.

File metadata

  • Download URL: goygram-0.7.47.tar.gz
  • Upload date:
  • Size: 78.1 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: maturin/1.15.0

File hashes

Hashes for goygram-0.7.47.tar.gz
Algorithm Hash digest
SHA256 8c8780fd0b43067e5614f8a3f81fe6b7bcdcd6a42cd23d32deb8b4ae14c3c11a
MD5 857891d85e30866c668633548c154fc0
BLAKE2b-256 49f92f7b6dc6479de86a288e050bcbddc245265fb1f9bb5ca4d5beb34ac35024

See more details on using hashes here.

File details

Details for the file goygram-0.7.47-cp311-abi3-win_amd64.whl.

File metadata

  • Download URL: goygram-0.7.47-cp311-abi3-win_amd64.whl
  • Upload date:
  • Size: 310.1 kB
  • Tags: CPython 3.11+, Windows x86-64
  • Uploaded using Trusted Publishing? No
  • Uploaded via: maturin/1.15.0

File hashes

Hashes for goygram-0.7.47-cp311-abi3-win_amd64.whl
Algorithm Hash digest
SHA256 324a98880c88a1742d7fea8d712989a8b2018776974aeb2cd5775de31e06ba9f
MD5 1f58bf8297d1080b5732a3043afefda5
BLAKE2b-256 cb0ee6f1649f88be7ae156fb56554d020314b50dde43a7af33adae7df51e1b80

See more details on using hashes here.

File details

Details for the file goygram-0.7.47-cp311-abi3-manylinux_2_34_x86_64.whl.

File metadata

File hashes

Hashes for goygram-0.7.47-cp311-abi3-manylinux_2_34_x86_64.whl
Algorithm Hash digest
SHA256 74400485fc0946eed115b4df984a5c8a07616356198012b31fc22fe65a8eee5a
MD5 d9afe7dfb9b1f3db56e64aac1281edb7
BLAKE2b-256 70840995b637d4b79d79c2e7009d090b59ab5c190fde4ac18a0edc5bea940e8a

See more details on using hashes here.

File details

Details for the file goygram-0.7.47-cp311-abi3-macosx_11_0_arm64.whl.

File metadata

File hashes

Hashes for goygram-0.7.47-cp311-abi3-macosx_11_0_arm64.whl
Algorithm Hash digest
SHA256 4ae5d2395d6ea608bb13cb6d02cc16ed05ba6d8db8a8c775135646df637610ce
MD5 1860c36da2b05216f405bc0a28797813
BLAKE2b-256 6112c7a9658c15680a3048225b93f65963ac51e9b064253d746f8ca19113f116

See more details on using hashes here.

Release history Release notifications | RSS feed

0.7.63

4 files

0.7.62

4 files

0.7.61

4 files

0.7.60

4 files

0.7.59

4 files

0.7.58

4 files

0.7.57

4 files

0.7.56

4 files

0.7.55

4 files

0.7.54

4 files

0.7.53

4 files

0.7.52

4 files

0.7.51

4 files

0.7.50

4 files

0.7.49

4 files

0.7.48

4 files

This release

0.7.47 This release

4 files

0.7.46

4 files

0.7.45

4 files

0.7.35

4 files

0.7.34

4 files

0.7.33

4 files

0.7.32

4 files

0.7.31

4 files

0.7.30

4 files

0.7.29

4 files

0.7.28

4 files

0.7.27

4 files

0.7.26

4 files

0.7.25

4 files

0.7.24

4 files

0.7.23

4 files

0.7.22

4 files

0.7.21

4 files

0.7.20

4 files

0.7.19

4 files

0.7.18

4 files

0.7.17

4 files

0.7.16

4 files

0.7.15

4 files

0.7.14

4 files

0.7.13

4 files

0.7.12

4 files

0.7.11

4 files

0.7.10

4 files

0.7.9

4 files

0.7.8

4 files

0.7.7

4 files

0.7.6

4 files

0.7.5

4 files

0.7.4

4 files

0.7.3

4 files

0.7.2

4 files

0.7.1

4 files

0.7.0

4 files

0.6.9

4 files

0.6.8

4 files

0.6.7

4 files

0.6.6

4 files

0.6.5

4 files

0.6.4

4 files

0.6.3

4 files

0.6.2

4 files

0.6.1

4 files

0.6.0

4 files

0.5.7

4 files

0.5.6

4 files

0.5.5

4 files

0.5.3

4 files

0.5.1

4 files

0.5.0

4 files

0.4.9

4 files

0.4.8

4 files

0.4.7

4 files

0.4.2

4 files

0.4.1

4 files

0.4.0

4 files

0.3.9

4 files

0.3.8

4 files

0.3.7

4 files

0.3.6

4 files

0.3.5

4 files

0.3.4

4 files

0.3.2

4 files

0.3.0

4 files

0.1.0

4 files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page