← 返回
未分类

GramGate

Telegram gateway for AI agents and automation. Use GramGate to access a real Telegram account over REST or MCP: read channels and history, search across chat...
Telegram gateway for AI agents and automation. Use GramGate to access a real Telegram account over REST or MCP: read channels and history, search across chat...
zakirovdmr zakirovdmr 来源
未分类 clawhub v0.1.1 1 版本 100000 Key: 无需
★ 0
Stars
📥 316
下载
💾 0
安装
1
版本
#latest

概述

GramGate

Telegram gateway for AI agents and automation.

Use GramGate when you need real Telegram account access via MTProto, not the limitations of the Bot API.

GitHub:

Why GramGate

Telegram bots cannot reliably do the things many serious workflows need: read subscribed channels, browse full history, search across chats, join by link, or click inline buttons as a normal user.

GramGate solves that by exposing a real Telegram account through:

  • REST API for scripts, automations, and HTTP tools
  • MCP server for Claude, GPT, IDE agents, and agent frameworks

Best Use Cases

Use this skill for:

  • channel research and monitoring, including subscribed-channel analysis and news tracking
  • content workflows, from research collection to drafting triggers and publishing operations
  • Telegram E2E testing, including bot flows that require real user interaction and button clicks
  • support, CRM, and notification flows inside Telegram
  • agent integrations, where an AI system needs structured Telegram read/write access

What It Can Do

GramGate exposes 40+ Telegram endpoints and supports workflows like:

  • reading dialogs, chats, channels, contacts, and full message history
  • searching within a chat or globally across Telegram
  • joining or leaving channels and groups
  • sending, editing, deleting, forwarding, copying, pinning, and reacting to messages
  • reading rich history with inline buttons and clicking those buttons
  • working with polls, media, contacts, locations, members, and live feed events

Typical local endpoints:

REST: http://127.0.0.1:18791
MCP:  http://127.0.0.1:18793/sse

Common REST Calls

Health and account

curl -s http://127.0.0.1:18791/health
curl -s http://127.0.0.1:18791/api/me
curl -s http://127.0.0.1:18791/api/dialogs?limit=30

Read chats and channels

curl -s -X POST http://127.0.0.1:18791/api/chat/info \
  -H 'Content-Type: application/json' \
  -d '{"chat_id":"@durov"}'

curl -s -X POST http://127.0.0.1:18791/api/chat/history \
  -H 'Content-Type: application/json' \
  -d '{"chat_id":"-1001234567890","limit":20}'

Search and join

curl -s -X POST http://127.0.0.1:18791/api/message/search \
  -H 'Content-Type: application/json' \
  -d '{"chat_id":"-1001234567890","query":"keyword","limit":20}'

curl -s -X POST http://127.0.0.1:18791/api/chat/join \
  -H 'Content-Type: application/json' \
  -d '{"link":"@channel_name"}'

Send a message

curl -s -X POST http://127.0.0.1:18791/api/message/send \
  -H 'Content-Type: application/json' \
  -d '{"recipient":"@username","text":"Hello from GramGate!"}'

Click an inline button

curl -s -X POST http://127.0.0.1:18791/api/button/click \
  -H 'Content-Type: application/json' \
  -d '{"chat_id":123456789,"message_id":42,"callback_data":"btn_action"}'

MCP Use

Use GramGate MCP when another agent should be able to:

  • list dialogs
  • inspect a chat or channel
  • fetch history
  • search messages
  • join channels
  • send messages
  • click Telegram inline buttons

Notes

  • chat_id may be numeric or @username
  • groups and channels usually use negative numeric ids
  • this is user-account access, so privacy and operational safety matter
  • prefer fewer, larger reads and avoid tight write loops
  • if exposing GramGate beyond localhost, use authentication and sane rate limits

Recommended Workflow

  1. Confirm the service is healthy.
  2. Discover the right chat or channel.
  3. Read history before taking action.
  4. Batch reads where possible.
  5. Write only when the task actually requires it.

Full Documentation

For setup, architecture, authentication, rate limiting, and the complete endpoint list, use the GitHub repo as the canonical reference:

版本历史

共 1 个版本

  • v0.1.1 当前
    2026-05-07 22:06 安全 安全

安全检测

腾讯云安全 (Keen)

安全,无风险
查看报告

腾讯云安全 (Sanbu)

安全,无风险
查看报告

🔗 相关推荐

ai-agent

Self-Improving + Proactive Agent

ivangdavila
自我反思+自我批评+自我学习+自组织记忆。智能体评估自身工作、发现错误并持续改进。
★ 1,414 📥 325,691
ai-agent

Skill Vetter

spclaudehome
AI智能体技能安全预审工具。安装ClawdHub、GitHub等来源技能前,检查风险信号、权限范围及可疑模式。
★ 1,244 📥 271,655
ai-agent

self-improving agent

pskoett
记录自身发现以实现自我改进的技能
★ 4,125 📥 880,564