Skip to content
Agents tracked: 444 Releases (7d): 644 Agent status: 6 with issuesSince last limit reset: Claude 31d · Codex 3d Updated Oct 23, 2026
AgentGid
✦ Ask

AI support agent for your customers: product recipe

For an agent that answers customers from your help center and past tickets, hands hard cases to a person, and works in your help desk. 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
Paid parts
no free tier listed: Freshdesk
If you outgrow the free tiers
about $78 / month
Freshdesk $19, Resend $20, LangSmith $39; billed by usage: Elasticsearch (Elastic Cloud), OpenRouter
Model, per 1,000 answers
about $0.6
Claude Haiku 5.5; $4.5 with Gemini 3.8 Flash; AgentGid estimate at ~4,000 input + 400 output tokens each

How it fits together

  1. Customer emails arrive in the help desk, which turns each one into a ticket.
  2. A webhook from the help desk sends the new ticket to your agent.
  3. The agent searches the knowledge base (help center articles, past solved tickets) and drafts a reply with the model.
  4. It posts the reply through the help desk API, or, when it is unsure or the customer asks for a person, assigns the ticket to a human with a note.
  5. Review a sample of answers every week and add missing articles.

Help desk and inbox

Freshdesk first pick Customer support software with ticketing, AI agents, and helpdesk features.
No free tier listed · paid from $19/mo · $49 per 100 sessions (AI Agent), $29/agent/month (AI Copilot), $2-$12/pass (day passes), $80 per 5,000 tasks,…
agent.md
Front Customer support platform with shared inbox, ticketing, and AI agents for omnichannel communication.
No free tier listed · paid from $25/mo · $0.05 per conversation (Autopilot AI add-on)
agent.md

Also fits: Intercom, Zendesk · all business apps with apis

Knowledge base search

Elasticsearch (Elastic Cloud) first pick
MCP
 Managed Elasticsearch for full-text, vector and hybrid search over indexed documents.
Open source · Serverless: Ingest as low as $0.14 per VCU per hour; Search as low as $0.09 per VCU per hour; Storage as low … · in 91 public agent repos
npm install @elastic/elasticsearch pip install elasticsearch agent.md
Zilliz Cloud
MCP
 Fully managed Milvus vector database with serverless and dedicated clusters.
Free: 5 GB storage, 2.5M vCUs per month, up to 5 collections · Dedicated: from $63 per million vectors/month (performance-optimized), from $16 (capacity-optimized), from $5… · in 108 public agent repos
npm install @zilliz/milvus2-sdk-node pip install pymilvus agent.md

Also fits: Qdrant Cloud, Chroma Cloud, LanceDB, Weaviate Cloud · all vector databases

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

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

Email

Resend first pick
MCP
 Developer email API for sending and receiving transactional and marketing email.
Free: 3,000 emails/month, 100/day, 3 domains · paid from $20/mo · Pro $20/month for 50,000 emails; extra emails $0.90 per 1,000 · in 426 public agent repos
npm install resend pip install resend agent.md

Also fits: Mailtrap Email API/SMTP, Brevo, Postmark, Elastic Email · all email sending

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
LangWatch
MCP
 LLM observability, evaluation and agent-simulation platform, open source and also hosted.
Free: Developer plan: 50k events/month, 14-day data access, 2 users, no credit card · Growth: EUR 29 per core-seat/month incl. 200k events, then EUR 5 per 100k events (prices shown in EUR only) · in 27 public agent repos
npm install langwatch pip install langwatch agent.md

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

Model

Default: fast and cheapClaude Haiku 5.5 · $0.1 input / $0.5 output per 1M tokens · or GPT-6 Luna, MiMo V2.6 Flash
Harder questionsGemini 3.8 Flash · $0.75 input / $3.75 output per 1M tokens · or Kimi K2.7 Code, DeepSeek V4 Pro
Hardest workKimi 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.

Or start from a ready platform

RasaFramework and platform for building voice and chat assistants that can be self-hosted: the open-source Rasa framework (Apache-2.0) and the open-core Rasa Pro /…
Free tier
Chatwoot CaptainChatwoot is an open-source omnichannel support desk (live chat, email, WhatsApp and social inboxes) and Captain is its AI agent that answers from help-center c…
Free tier · from $19/mo
DecagonPlatform for deploying AI agents that resolve customer support conversations over chat, email and voice, with workflows written as natural-language Agent Opera…
BotpressPlatform for building and deploying LLM-based support agents with a hosted studio, a conversation-metered pricing model and a public MIT-licensed monorepo of S…
Free tier · from $150/mo

Libraries

AI SDKVercel's unified SDK for building AI apps with OpenAI, Anthropic, Google and more.
npm install ai · Apache-2.0 · Gid Score 80
LlamaIndexFramework to build RAG applications connecting LLMs to your custom data and documents.
pip install llama-index · MIT · Gid Score 63
LangGraphFramework for building stateful, multi-actor applications with LLMs.
pip install langgraph · MIT · Gid Score 78
Pydantic AIAI agent framework built with Pydantic for structured, typed agent development.
pip install pydantic-ai · MIT · Gid Score 62

Coding agent to build it

OpenAI Codex Claude Code OpenCode Pi Compare coding agents

Before you ship

  • Check what the help desk's own AI add-ons cost before building: some help desks bill AI answers separately.
  • Never let the agent promise refunds or account changes; route those to a person.
  • 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: AI support agent for your customers (AgentGid)

For an agent that answers customers from your help center and past tickets, hands hard cases to a person, and works in your help desk.

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: these first picks have no free tier listed: Freshdesk.
- If you outgrow the free tiers, the lowest paid plans of the first picks are Freshdesk $19, Resend $20, LangSmith $39 (together about $78/month); usage beyond a plan's credits is billed on top.
- Billed by usage (grows with traffic): Elasticsearch (Elastic Cloud), OpenRouter.
- Model, default: fast and cheap: about $0.6 per 1,000 answers with Claude Haiku 5.5 (AgentGid estimate at ~4,000 input + 400 output tokens each).
- Model, harder questions: about $4.5 per 1,000 answers with Gemini 3.8 Flash (AgentGid estimate at ~4,000 input + 400 output tokens each).
- Prices of the picks below were checked on the official pages between 2026-10-10 and 2026-10-20.

## How it fits together
1. Customer emails arrive in the help desk, which turns each one into a ticket.
2. A webhook from the help desk sends the new ticket to your agent.
3. The agent searches the knowledge base (help center articles, past solved tickets) and drafts a reply with the model.
4. It posts the reply through the help desk API, or, when it is unsure or the customer asks for a person, assigns the ticket to a human with a note.
5. Review a sample of answers every week and add missing articles.

## Help desk and inbox
- Freshdesk: no free tier listed · paid from $19/mo · usage: $49 per 100 sessions (AI Agent), $29/agent/month (AI Copilot), $2-$12/pass (day passes), $80 per 5,000 tasks, $15 per 1,000 actions · card https://agentgid.com/apis/freshdesk/agent.md
- Front: no free tier listed · paid from $25/mo · usage: $0.05 per conversation (Autopilot AI add-on) · card https://agentgid.com/apis/front/agent.md
- Also fits: Intercom, Zendesk (https://agentgid.com/apis/business-apps/)

## Knowledge base search
- Elasticsearch (Elastic Cloud): open source · usage: Serverless: Ingest as low as $0.14 per VCU per hour; Search as low as $0.09 per VCU per hour; Storage as low as $0.047 per GB retained per month · official MCP https://github.com/elastic/mcp-server-elasticsearch · npm @elastic/elasticsearch · pip elasticsearch · card https://agentgid.com/apis/elastic-cloud/agent.md
- Zilliz Cloud: free: 5 GB storage, 2.5M vCUs per month, up to 5 collections · usage: Dedicated: from $63 per million vectors/month (performance-optimized), from $16 (capacity-optimized), from $5 (tiered-storage); Standard Dedicated from $126/GB/month · official MCP https://github.com/zilliztech/mcp-server-milvus · npm @zilliz/milvus2-sdk-node · pip pymilvus · card https://agentgid.com/apis/zilliz-cloud/agent.md
- Also fits: Qdrant Cloud, Chroma Cloud, LanceDB, Weaviate Cloud (https://agentgid.com/apis/vector-db/)

## 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
- Also fits: LiteLLM, Vercel AI Gateway, Cloudflare AI Gateway, Portkey (https://agentgid.com/apis/llm-gateway/)

## Email
- Resend: free: 3,000 emails/month, 100/day, 3 domains · paid from $20/mo · usage: Pro $20/month for 50,000 emails; extra emails $0.90 per 1,000 · official MCP resend-mcp · npm resend · pip resend · card https://agentgid.com/apis/resend/agent.md
- Also fits: Mailtrap Email API/SMTP, Brevo, Postmark, Elastic Email (https://agentgid.com/apis/email/)

## 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
- LangWatch: free: Developer plan: 50k events/month, 14-day data access, 2 users, no credit card · usage: Growth: EUR 29 per core-seat/month incl. 200k events, then EUR 5 per 100k events (prices shown in EUR only) · official MCP @langwatch/mcp-server · npm langwatch · pip langwatch · card https://agentgid.com/apis/langwatch/agent.md
- Also fits: W&B Weave, Ragas, Pydantic Logfire, AgentOps (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

## Or start from a ready platform
- Rasa: free tier · https://agentgid.com/agents/rasa/agent.md
- Chatwoot Captain: free tier, from $19/mo · https://agentgid.com/agents/chatwoot-captain/agent.md
- Decagon: see card · https://agentgid.com/agents/decagon/agent.md
- Botpress: free tier, from $150/mo · https://agentgid.com/agents/botpress/agent.md

## Libraries
- `npm install ai`: AI SDK (Apache-2.0) · https://agentgid.com/tools/ai/agent.md
- `pip install llama-index`: LlamaIndex (MIT) · https://agentgid.com/tools/llama-index/agent.md
- `pip install langgraph`: LangGraph (MIT) · https://agentgid.com/tools/langgraph/agent.md
- `pip install pydantic-ai`: Pydantic AI (MIT) · https://agentgid.com/tools/pydantic-ai/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

## Before you ship
- Check what the help desk's own AI add-ons cost before building: some help desks bill AI answers separately.
- Never let the agent promise refunds or account changes; route those to a person.
- 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/support-agent/ · 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 Web research and scraping agent Telegram bot with AI 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 2026-10-10 and 2026-10-20; check them before you commit.