Telegram bot with AI: product recipe

For a Telegram bot that chats with a model, remembers users, and runs around the clock. Every service below is from AgentGid's catalog, with the price and free tier from its official page; the first pick in each part ranks best on free tier, official MCP server, real use and price.

agent.md

What it costs

To start

$0 / month

every first pick has a free tier or is open source (limits apply)

If you outgrow the free tiers

about $78 / month

Supermemory $19, Vercel $20, LangSmith $39; billed by usage: Telegram Bot API, OpenRouter, Neon

Model, per 1,000 messages

about $0.35

Claude Haiku 5.5; $2.62 with Gemini 3.8 Flash; AgentGid estimate at ~2,000 input + 300 output tokens each

How it fits together

Create the bot with @BotFather and keep its token on the server.
Receive updates by webhook (serverless hosting works) or long polling (needs an always-on worker).
Key memory and history by the Telegram user id; send recent context to the model through the gateway.
Rate-limit per user and cap spending at the gateway.

Telegram Bot API

Telegram Bot API first pick	HTTP API for building Telegram bots that send and receive messages in chats and groups. Free: Bots are able to message their users at no cost (about 30 messages per second broadcast limit) · Paid broadcasts: 0.1 Stars per message over the free 30 per second (up to 1000 messages per second) agent.md

LLM gateway (one key, many models)

OpenRouter first pick MCP	One OpenAI-compatible API for 500+ models across 80+ providers with routing and fallbacks. Free: 25+ free models, 50 requests/day · Provider list price per token (no markup); fee on credit purchases: 5.5% Standard, 8% Business · in 29 public agent repos `npm install @openrouter/sdk` `pip install openrouter` agent.md
LiteLLM	Open-source proxy and SDK that calls 100+ LLMs in OpenAI format with keys, budgets and logging. Free: Open-source gateway free to self-host · in 657 public agent repos `pip install litellm` agent.md

Also fits: Vercel AI Gateway, Cloudflare AI Gateway, Portkey, Helicone · all llm gateways and routers

Memory for the bot

Supermemory first pick MCP	Memory and retrieval API that builds per-user memory graphs from text, files and connectors. Free: $5 of credits/month · paid from $19/mo · Memory $5 per 1M SM tokens (plain text), $10 (rich); SuperRAG $1-$2 per 1M tokens; search $5 per 1M queries · in 21 public agent repos `npm install supermemory` `pip install supermemory` agent.md
Mem0 MCP	Memory layer that extracts, stores and retrieves user and agent memories for LLM apps. Free: Hobby: 10,000 add requests and 1,000 retrieval requests/month · paid from $19/mo · Starter $19/month: 50,000 adds + 5,000 retrievals; Pro $249/month: 500,000 adds + 50,000 retrievals · in 359 public agent repos `npm install mem0ai` `pip install mem0ai` agent.md

Also fits: Letta, Zep · all memory for agents

Database

Neon first pick MCP	Managed Postgres database with autoscaling compute, storage, and backend services. Free: 1 GB storage, 100 CU-hours compute/month per project, 100 projects, 5 GB Object Storage, 1M function invocations · $0.106 per CU-hr (Launch), $0.222 per CU-hr (Scale); $0.35 per GB-month storage; $0.023 per GB-month Object S… · in 306 public agent repos `npm install @neondatabase/serverless` agent.md
Upstash Redis	Managed Redis database with global replication and vector search capabilities. Free: 250 MB data, 500K commands/month · paid from $10/mo · $0.20 per 100K commands, $0.25 per GB storage (after 1GB free), $0.03 per GB bandwidth (after 200GB free) · in 218 public agent repos `npm install @upstash/redis` agent.md

Also fits: Redis, Convex, Firebase, Appwrite · all databases and backends

Hosting (webhook or always-on worker)

Vercel first pick MCP	Hosting for web apps and Next.js with a CDN and serverless functions. Free: Hobby: $0/month · paid from $20/mo · in 102 public agent repos `npm install vercel` agent.md
Azure Functions MCP server page	Run event-driven serverless functions on Azure, billed per execution and GB-second. Free: Flex Consumption: 250,000 executions and 100,000 GB-s per month (Consumption plan: 1 million executions and 400,000 GB-s) · Flex Consumption on-demand (East US): $0.000026 per GB-s, $0.40 per million executions; Consumption plan: $0.… · in 38 public agent repos `npm install @azure/functions` `pip install azure-functions` agent.md

Vercel first pick

MCP

Hosting for web apps and Next.js with a CDN and serverless functions.

Free: Hobby: $0/month · paid from $20/mo · in 102 public agent repos

npm install vercel agent.md

Azure Functions

MCP server page

Run event-driven serverless functions on Azure, billed per execution and GB-second.

Free: Flex Consumption: 250,000 executions and 100,000 GB-s per month (Consumption plan: 1 million executions and 400,000 GB-s) · Flex Consumption on-demand (East US): $0.000026 per GB-s, $0.40 per million executions; Consumption plan: $0.… · in 38 public agent repos

npm install @azure/functions pip install azure-functions agent.md

Also fits: DigitalOcean App Platform, Netlify, Heroku, Cloudflare Workers · all hosting and deployment

Traces and evals

LangSmith first pick MCP	Tracing, evaluation and monitoring of LLM apps and agents, including prompt and dataset management. Free: Developer plan: 1 seat, up to 5k base traces / mo, then pay-as-you-go · paid from $39/mo · Plus: $39 / seat per month incl. up to 10k base traces / mo; trace overage is pay-as-you-go (rate not listed … · in 678 public agent repos `npm install langsmith` `pip install langsmith` agent.md

Also fits: LangWatch, W&B Weave, Ragas, Pydantic Logfire · all llm observability and evals

Model

Default: fast and cheap	Claude Haiku 5.5 · $0.1 input / $0.5 output per 1M tokens · or GPT-6 Luna, MiMo V2.6 Flash
Harder questions	Gemini 3.8 Flash · $0.75 input / $3.75 output per 1M tokens · or Kimi K2.7 Code, DeepSeek V4 Pro
Hardest work	Kimi K3 · $3 input / $15 output per 1M tokens · or Claude Opus 5.5, Claude Fable 5.1

The cheapest model of each price tier that can call tools, from model prices. Try the default first and move up only for the questions it gets wrong.

Libraries

aiogram	Asynchronous framework for building Telegram bots with modern async/await patterns. `pip install aiogram` · MIT · Gid Score 50
python-telegram-bot	Python wrapper library for building Telegram bots with the Telegram Bot API. `pip install python-telegram-bot` · LGPL-3.0-only · Gid Score 64
grammY	Telegram Bot Framework for building bots with JavaScript/TypeScript. `npm install grammy` · MIT · Gid Score 51
Telegraf	Modern Telegram Bot Framework for building bots with intuitive APIs. `npm install telegraf` · MIT · Gid Score 40

Coding agent to build it

OpenAI Codex Claude Code OpenCode Pi Compare coding agents

Also on AgentGid

Telegram news bot kit

Before you ship

Keep API keys on the server, never in the browser or app; set spending limits where the provider offers them.
Read each service's terms for your use: data retention, commercial use, rate limits.
Log model calls (traces) and test with your own real inputs before launch.

The brief

# Product recipe: Telegram bot with AI (AgentGid)

For a Telegram bot that chats with a model, remembers users, and runs around the clock.

Parts were picked from AgentGid's catalog by rules: free tier, official MCP server, use in public projects built with AI agents, downloads, price. Prices are list prices from the official pages (see each card for the check date); confirm before you commit.

## What it costs
- Start: $0 a month on the free tiers of the first picks (their limits apply).
- If you outgrow the free tiers, the lowest paid plans of the first picks are Supermemory $19, Vercel $20, LangSmith $39 (together about $78/month); usage beyond a plan's credits is billed on top.
- Billed by usage (grows with traffic): Telegram Bot API, OpenRouter, Neon.
- Model, default: fast and cheap: about $0.35 per 1,000 messages with Claude Haiku 5.5 (AgentGid estimate at ~2,000 input + 300 output tokens each).
- Model, harder questions: about $2.62 per 1,000 messages with Gemini 3.8 Flash (AgentGid estimate at ~2,000 input + 300 output tokens each).
- Prices of the picks below were checked on the official pages between 2026-10-09 and 2026-10-22.

## How it fits together
1. Create the bot with @BotFather and keep its token on the server.
2. Receive updates by webhook (serverless hosting works) or long polling (needs an always-on worker).
3. Key memory and history by the Telegram user id; send recent context to the model through the gateway.
4. Rate-limit per user and cap spending at the gateway.

## Telegram Bot API
- Telegram Bot API: free: Bots are able to message their users at no cost (about 30 messages per second broadcast limit) · usage: Paid broadcasts: 0.1 Stars per message over the free 30 per second (up to 1000 messages per second) · card https://agentgid.com/apis/telegram-bot-api/agent.md

## LLM gateway (one key, many models)
- OpenRouter: free: 25+ free models, 50 requests/day · usage: Provider list price per token (no markup); fee on credit purchases: 5.5% Standard, 8% Business · official MCP https://mcp.openrouter.ai/mcp · npm @openrouter/sdk · pip openrouter · card https://agentgid.com/apis/openrouter/agent.md
- LiteLLM: free: Open-source gateway free to self-host · pip litellm · card https://agentgid.com/apis/litellm/agent.md
- Also fits: Vercel AI Gateway, Cloudflare AI Gateway, Portkey, Helicone (https://agentgid.com/apis/llm-gateway/)

## Memory for the bot
- Supermemory: free: $5 of credits/month · paid from $19/mo · usage: Memory $5 per 1M SM tokens (plain text), $10 (rich); SuperRAG $1-$2 per 1M tokens; search $5 per 1M queries · official MCP https://mcp.supermemory.ai/mcp · npm supermemory · pip supermemory · card https://agentgid.com/apis/supermemory/agent.md
- Mem0: free: Hobby: 10,000 add requests and 1,000 retrieval requests/month · paid from $19/mo · usage: Starter $19/month: 50,000 adds + 5,000 retrievals; Pro $249/month: 500,000 adds + 50,000 retrievals · official MCP https://mcp.mem0.ai/mcp · npm mem0ai · pip mem0ai · card https://agentgid.com/apis/mem0/agent.md
- Also fits: Letta, Zep (https://agentgid.com/apis/memory/)

## Database
- Neon: free: 1 GB storage, 100 CU-hours compute/month per project, 100 projects, 5 GB Object Storage, 1M function invocations · usage: $0.106 per CU-hr (Launch), $0.222 per CU-hr (Scale); $0.35 per GB-month storage; $0.023 per GB-month Object Storage; $0.60 per 1M invocations · official MCP · npm @neondatabase/serverless · card https://agentgid.com/apis/neon/agent.md
- Upstash Redis: free: 250 MB data, 500K commands/month · paid from $10/mo · usage: $0.20 per 100K commands, $0.25 per GB storage (after 1GB free), $0.03 per GB bandwidth (after 200GB free) · npm @upstash/redis · card https://agentgid.com/apis/upstash/agent.md
- Also fits: Redis, Convex, Firebase, Appwrite (https://agentgid.com/apis/backend-db/)

## Hosting (webhook or always-on worker)
- Vercel: free: Hobby: $0/month · paid from $20/mo · official MCP https://mcp.vercel.com · npm vercel · card https://agentgid.com/apis/vercel/agent.md
- Azure Functions: free: Flex Consumption: 250,000 executions and 100,000 GB-s per month (Consumption plan: 1 million executions and 400,000 GB-s) · usage: Flex Consumption on-demand (East US): $0.000026 per GB-s, $0.40 per million executions; Consumption plan: $0.000016 per GB-s, $0.20 per million executions · official MCP @azure/mcp · npm @azure/functions · pip azure-functions · card https://agentgid.com/apis/azure-functions/agent.md · MCP server card https://agentgid.com/mcp/servers/com-microsoft-azure/agent.md
- Also fits: DigitalOcean App Platform, Netlify, Heroku, Cloudflare Workers (https://agentgid.com/apis/hosting/)

## Traces and evals
- LangSmith: free: Developer plan: 1 seat, up to 5k base traces / mo, then pay-as-you-go · paid from $39/mo · usage: Plus: $39 / seat per month incl. up to 10k base traces / mo; trace overage is pay-as-you-go (rate not listed on pricing page); 1 LSU = $1 (Deployment/Fleet/Sandbox usage) · official MCP https://docs.langchain.com/langsmith/langsmith-mcp-server · npm langsmith · pip langsmith · card https://agentgid.com/apis/langsmith/agent.md
- Also fits: LangWatch, W&B Weave, Ragas, Pydantic Logfire (https://agentgid.com/apis/observability/)

## Model (all of these can call tools; prices per 1M tokens)
- Default: fast and cheap: Claude Haiku 5.5 ($0.1 in / $0.5 out) · https://agentgid.com/models/claude-haiku-5-5/agent.md · or GPT-6 Luna, MiMo V2.6 Flash
- Harder questions: Gemini 3.8 Flash ($0.75 in / $3.75 out) · https://agentgid.com/models/gemini-3-8-flash/agent.md · or Kimi K2.7 Code, DeepSeek V4 Pro
- Hardest work: Kimi K3 ($3 in / $15 out) · https://agentgid.com/models/kimi-k3/agent.md · or Claude Opus 5.5, Claude Fable 5.1

## Libraries
- `pip install aiogram`: aiogram (MIT) · https://agentgid.com/tools/aiogram/agent.md
- `pip install python-telegram-bot`: python-telegram-bot (LGPL-3.0-only) · https://agentgid.com/tools/python-telegram-bot/agent.md
- `npm install grammy`: grammY (MIT) · https://agentgid.com/tools/grammy/agent.md
- `npm install telegraf`: Telegraf (MIT) · https://agentgid.com/tools/telegraf/agent.md

## Coding agent to build it
- OpenAI Codex: free tier, from $8/mo · https://agentgid.com/agents/openai-codex/agent.md
- Claude Code: from $20/mo · https://agentgid.com/agents/claude-code/agent.md
- OpenCode: free tier, from $10/mo · https://agentgid.com/agents/opencode/agent.md

## Also on AgentGid
- Telegram news bot kit: https://agentgid.com/build/kits/news-bot/

## Before you ship
- Keep API keys on the server, never in the client; set spending limits where the provider offers them.
- Read each service's terms for your use (data retention, commercial use, rate limits).
- Log model calls (traces) and test with your own real inputs before launch.

Recipe page: https://agentgid.com/build/recipes/telegram-ai-bot/ · all product recipes: https://agentgid.com/build/recipes/agent.md

Other recipes

AI assistant over your documents (RAG) Voice agent (phone or web) AI chatbot for a website or app AI support agent for your customers Web research and scraping agent Data analyst agent Solana trading or alert bot Video from a prompt Design recipes

Picked by rules from AgentGid's catalog, not paid placement. Prices are list prices from the official pages, checked between 2026-10-09 and 2026-10-22; check them before you commit.