pydantic-ai demos

README.md

@@ -45,6 +45,11 @@ demonstrate basic system features, integration APIs, and best practices.
 | [Jira assignee from Google Calendar](./jira_google_calendar/assignee_from_schedule/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=jira_google_calendar/assignee_from_schedule) | Set assignee in Jira ticket to the person currently on-call | jira, googlecalendar |
 | [Create calendar due date event for Jira ticket](./jira_google_calendar/deadline_to_event/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=jira_google_calendar/deadline_to_event) | When a new Jira issue is created, the workflow automatically generates a Google Calendar event with a deadline | googlecalendar, jira |
 | [Leash - Incident Management & On-Call Escalation](./leash/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=leash) | Automated incident management system with on-call rotation, escalation workflows, and multi-channel notifications | googlesheets, slack, twilio, gmail |
+| [Ask - Slack Q&A Bot](./pydantic/ask/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic/ask) | Simple one-shot question answering bot for Slack using Pydantic AI | slack |
+| [Chat - Conversational Slack Bot](./pydantic/chat/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic/chat) | Multi-turn conversation bot for Slack with persistent history using Pydantic AI | slack |
+| [Chat with UI - Web-Based Conversational AI](./pydantic/chat_with_ui/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic/chat_with_ui) | Browser-based chat interface with persistent conversation history using Pydantic AI |  |
+| [D&D Party Chat - AI-Powered Multiplayer RPG](./pydantic/dnd/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic/dnd) | Web-based D&D game where you adventure with AI teammates, powered by multiple LLM providers | pydanticgw |
+| [Gamble - AI Casino Bot](./pydantic/gamble/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic/gamble) | Interactive roulette and blackjack games powered by AI tool calling with Pydantic AI | slack |
 | [Quickstart](./quickstart/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=quickstart) | Sample for quickstart |  |
 | [AWS Health monitor](./reliability/aws_health_monitor/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=reliability/aws_health_monitor) | Announce AWS Health events in Slack channels, based on resource ownership data in a Google Sheet | aws, slack, googlesheets |
 | [Missing Jira events monitor](./reliability/missing_jira_events_monitor/)<br/>[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=reliability/missing_jira_events_monitor) | Send Slack alerts when AutoKitteh doesn't receive certain Jira events in time | jira, slack |

pydantic/ask/README.md

@@ -0,0 +1,183 @@
+---
+title: Ask - Slack Q&A Bot
+description: Simple one-shot question answering bot for Slack using Pydantic AI
+integrations: ["slack"]
+categories: ["AI", "Samples"]
+tags:
+  [
+    "slack_bot",
+    "pydantic_ai",
+    "question_answering",
+    "stateless",
+    "single_turn",
+  ]
+---
+
+# Ask - Slack Q&A Bot 🤖
+
+[![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic-ask)
+
+A simple Slack bot that answers questions using Pydantic AI. Send a message starting with `!ask` and get an instant AI-powered response in a thread. Each question is treated independently without conversation history.
+
+### What AutoKitteh Provides
+
+- **Slack Integration**: Seamless Slack event handling and message posting
+- **Event Filtering**: Automatic routing of `!ask` commands to the handler
+- **Connection Management**: Secure credential storage for Slack and AI providers
+- **Reliable Execution**: Guaranteed message processing and response delivery
+
+## Features
+
+- **One-Shot Q&A**: Each question is independent with no conversation memory
+- **Threaded Responses**: Replies appear as threads to keep channels clean
+- **Configurable Models**: Switch between any Pydantic AI-compatible model
+- **Simple Integration**: Just add the bot to your Slack workspace and start asking
+
+## How It Works
+
+1. **User posts message**: `!ask What is the capital of France?`
+2. **AutoKitteh triggers**: Filters and routes the message to the handler
+3. **AI processes**: Pydantic AI agent generates a concise response
+4. **Bot replies**: Response appears in a thread attached to the original message
+
+## Architecture
+
+```
+Slack Message (!ask) → AutoKitteh Event Trigger → Pydantic AI Agent → Slack Thread Reply
+```
+
+### Key Components
+
+- **`handlers.py`**: Event handler that processes Slack messages and generates AI responses
+- **`autokitteh.yaml`**: AutoKitteh project configuration with Slack trigger and event filter
+
+## Setup
+
+### Prerequisites
+
+- AutoKitteh account (cloud or self-hosted)
+- Slack workspace where you can install apps
+- API key for your chosen AI provider (Anthropic, OpenAI, etc.)
+
+### Installation
+
+1. Start using AutoKitteh Cloud:
+
+   [![Start with AutoKitteh](https://autokitteh.com/assets/autokitteh-badge.svg)](https://app.autokitteh.cloud/template?template-name=pydantic-ask)
+
+2. Configure your API key:
+   - Navigate to the project configuration in AutoKitteh UI
+   - Go to the **Variables** tab
+   - Set `ANTHROPIC_API_KEY` (or your provider's key)
+   - Optionally change `MODEL_NAME` (default: `anthropic:claude-sonnet-4-0`)
+
+3. Initialize the Slack connection:
+   - Go to the **Connections** tab
+   - Initialize the `slack` connection
+   - Follow the OAuth flow to authorize your Slack workspace
+
+4. Invite the bot to a channel:
+   - In Slack, type `/invite @AutoKitteh` in any channel
+   - Or add it from the channel details
+
+## Usage
+
+### Asking Questions
+
+Simply type `!ask` followed by your question in any channel where the bot is present:
+
+```
+!ask What is the speed of light?
+!ask Explain quantum entanglement in simple terms
+!ask Who wrote "The Great Gatsby"?
+```
+
+The bot will respond in a thread with a concise answer.
+
+### Changing the AI Model
+
+Update the `MODEL_NAME` variable in `autokitteh.yaml` or in the AutoKitteh UI:
+
+```yaml
+vars:
+  - name: MODEL_NAME
+    value: "anthropic:claude-sonnet-4-0"  # or "openai:gpt-4o", etc.
+```
+
+Supported model formats:
+- Anthropic: `anthropic:claude-sonnet-4-0`
+- OpenAI: `openai:gpt-4o`
+- Any other Pydantic AI-compatible provider
+
+## Technical Details
+
+### Pydantic AI Integration
+
+The project uses [Pydantic AI](https://ai.pydantic.dev/) for AI interactions:
+
+- **Stateless Agent**: No conversation history - each question is independent
+- **Concise Instructions**: Agent is configured to respond in one sentence
+- **Synchronous Processing**: Uses `run_sync()` for simple request/response flow
+
+### Event Filtering
+
+The AutoKitteh trigger uses a filter to only process relevant messages:
+
+```python
+filter: "data.thread_ts == '' && data.text.startsWith('!ask')"
+```
+
+This ensures:
+- Only top-level messages are processed (not thread replies)
+- Only messages starting with `!ask` trigger the bot
+
+### Response Format
+
+Responses are posted as threaded replies with the format:
+
+```
+`<model-name>` says:
+```<answer>```
+```
+
+This keeps the main channel clean while providing context about which model generated the response.
+
+## Comparison with Other Samples
+
+| Sample | Conversation History | UI | Use Case |
+|--------|---------------------|-----|----------|
+| **ask** | ❌ No | Slack | Quick one-off questions |
+| **chat** | ✅ Yes | Slack | Multi-turn conversations |
+| **chat_with_ui** | ✅ Yes | Web | Browser-based chat |
+| **gamble** | ✅ Yes | Slack | Interactive AI games |
+
+## Development
+
+Run type checking and validation locally:
+
+```bash
+ak make
+```
+
+Deploy from the command line:
+
+1. **Install the CLI**: https://docs.autokitteh.com/get_started/install
+2. **Authenticate**: `ak auth login`
+3. **Deploy**: `ak deploy`
+4. **Initialize connections**: Log in to https://autokitteh.cloud and set up Slack
+5. **Start asking**: Use `!ask` in Slack
+
+## Customization Ideas
+
+- **Change response style**: Modify the agent's instructions for different tones
+- **Add context**: Pass additional information to the agent (company docs, user data, etc.)
+- **Multi-language**: Detect language and respond accordingly
+- **Rate limiting**: Track usage per user or channel
+- **Custom commands**: Add more `!` commands for different behaviors
+
+## Known Limitations
+
+- No conversation memory across questions
+- Single sentence responses (by design)
+- Requires the bot to be in the channel
+- No support for direct messages (DMs) - only channels

pydantic/ask/autokitteh.yaml

@@ -0,0 +1,24 @@
+version: v2
+
+project:
+  name: pydantic_ask
+
+  vars:
+    - name: ANTHROPIC_API_KEY
+      value: ""
+      secret: true
+    - name: MODEL_NAME
+      value: "claude-sonnet-4-0"
+
+  connections:
+    - name: slack
+      integration: slack
+    - name: anthropic
+      integration: anthropic
+
+  triggers:
+    - name: slack_message
+      connection: slack
+      event_type: message
+      filter: "data.thread_ts == '' && data.text.startsWith('!ask')"
+      call: handlers.py:on_slack_message

pydantic/ask/handlers.py

@@ -0,0 +1,80 @@
+"""Slack Q&A Bot - Simple one-shot question answering using Pydantic AI
+
+This module implements a stateless Slack bot that answers questions using AI.
+Each message is treated independently without conversation history.
+
+Architecture:
+    Slack Message (!ask) → AutoKitteh Trigger → Pydantic AI Agent → Slack Reply
+
+Key Features:
+    - One-shot Q&A (no conversation history)
+    - Configurable AI model via environment variable
+    - Threaded replies in Slack
+    - Synchronous processing
+"""
+
+from os import getenv
+
+from pydantic_ai import Agent
+from pydantic_ai.models.anthropic import AnthropicModel
+
+from autokitteh import Event
+from autokitteh.pydantic import anthropic_pydantic_ai_provider
+from autokitteh.slack import slack_client
+
+
+# Initialize Slack client using AutoKitteh connection
+_slack = slack_client("slack")
+
+# AI model configuration
+_MODEL_NAME = getenv("MODEL_NAME", "claude-sonnet-4-0")
+
+# Create Anthropic model with AutoKitteh provider
+model = AnthropicModel(
+    _MODEL_NAME,
+    provider=anthropic_pydantic_ai_provider("anthropic")
+)
+
+# Create Pydantic AI agent with the configured model
+# This agent is stateless - each run is independent
+agent = Agent(
+    model=model,
+    instructions="Be concise, reply with one sentence.",
+)
+
+
+def on_slack_message(event: Event) -> None:
+    """Handle incoming Slack messages starting with !ask command.
+
+    Triggered by Slack messages matching the filter in autokitteh.yaml:
+    - Must not be in a thread (thread_ts == '')
+    - Must start with '!ask'
+
+    Args:
+        event: AutoKitteh event containing Slack message data
+
+    Process:
+        1. Extract question text (remove !ask prefix)
+        2. Send to Pydantic AI agent for processing
+        3. Post response back to Slack in a thread
+
+    Note:
+        No conversation history is maintained - each question is independent.
+    """
+    data = event.data
+    # Remove the !ask command prefix and any leading/trailing whitespace
+    q = data.text.removeprefix("!ask").strip()
+
+    print(f"Q: {q}")
+
+    # Run agent synchronously - no message history passed
+    a = agent.run_sync(q).output
+
+    print(f"A: {a}")
+
+    # Post response in a thread attached to the original message
+    _slack.chat_postMessage(
+        channel=event.data.channel,
+        text=f"```{a}```",
+        thread_ts=event.data.ts,  # Creates a thread reply
+    )