feat: Fireworks as inference option

python/ai-data-analyst/.env.example

@@ -1,6 +1,7 @@
 DAYTONA_API_KEY=
 ANTHROPIC_API_KEY=
 OPENAI_API_KEY=
+FIREWORKS_AI_API_KEY=
 MISTRAL_API_KEY=
 DEEPSEEK_API_KEY=
-OPENROUTER_API_KEY=
+OPENROUTER_API_KEY=

python/ai-data-analyst/README.md

@@ -33,6 +33,7 @@ To run this example, you need to set the following environment variables:
 
 - `ANTHROPIC_API_KEY`: Required if using Anthropic models (default)
 - `OPENAI_API_KEY`: Required if using OpenAI models
+- `FIREWORKS_AI_API_KEY`: Required if using Fireworks AI models
 - `MISTRAL_API_KEY`: Required if using Mistral AI models
 - `DEEPSEEK_API_KEY`: Required if using DeepSeek models
 - `OPENROUTER_API_KEY`: Required if using OpenRouter models
@@ -99,6 +100,7 @@ The coding model is used for high accuracy code generation, and the summary mode
 Other suggested models include:
 
 - `openai/gpt-5.1`
+- `fireworks_ai/accounts/fireworks/models/kimi-k2p6`
 - `mistral/mistral-large-latest`
 - `deepseek/deepseek-chat`
 - `openrouter/moonshotai/kimi-k2`

python/langchain/data-analysis/.env.example

@@ -1,2 +1,7 @@
 DAYTONA_API_KEY=
-ANTHROPIC_API_KEY=
+# Model in "provider:model" format; defaults to the Anthropic model below.
+# MODEL=anthropic:claude-sonnet-4-5-20250929
+# Set the API key matching your chosen provider:
+ANTHROPIC_API_KEY=
+# FIREWORKS_API_KEY=
+# OPENAI_API_KEY=

python/langchain/data-analysis/README.md

@@ -27,7 +27,11 @@ In this example, the agent analyzes a vehicle valuations dataset to understand h
 To run this example, you need to set the following environment variables:
 
 - `DAYTONA_API_KEY`: Required for access to Daytona sandboxes. Get it from [Daytona Dashboard](https://app.daytona.io/dashboard/keys)
-- `ANTHROPIC_API_KEY`: Required for Claude AI model access. Get it from [Anthropic Console](https://console.anthropic.com/)
+- `MODEL` (optional): The model in `provider:model` format. Defaults to `anthropic:claude-sonnet-4-5-20250929`
+- The API key matching your chosen provider:
+  - `ANTHROPIC_API_KEY`: Required if using Anthropic models (default)
+  - `FIREWORKS_API_KEY`: Required if using Fireworks AI models
+  - `OPENAI_API_KEY`: Required if using OpenAI models
 
 See the `.env.example` file for the exact structure. Copy `.env.example` to `.env` and fill in your API keys before running.
 
@@ -70,6 +74,17 @@ Before proceeding, complete the following steps:
 
 ## Configuration
 
+- **LLM Provider:** The model is set via the `MODEL` environment variable in `provider:model` format and works with any [provider supported by LangChain](https://docs.langchain.com/oss/python/integrations/chat). Install the provider's integration package and set its API key. For example, to use [Fireworks AI](https://fireworks.ai/):
+
+  ```bash
+  pip install -U langchain-fireworks
+  ```
+
+  ```env
+  MODEL=fireworks:accounts/fireworks/models/kimi-k2p6
+  FIREWORKS_API_KEY=your-fireworks-api-key
+  ```
+
 - **Analysis Prompt:** The main prompt is configured in the `agent.invoke()` call inside `data_analysis.py`. You can modify this prompt to analyze different aspects of the data or try different visualization types.
 
 - **Dataset Description:** When uploading the dataset, provide a clear description of the columns and data cleaning instructions to help the agent understand how to process the data.

python/langchain/data-analysis/data_analysis.py

@@ -4,10 +4,11 @@
 """LangChain data analysis example using Daytona sandboxes."""
 
 import base64
+import os
 
 from dotenv import load_dotenv
 from langchain.agents import create_agent  # pylint: disable=import-error
-from langchain_anthropic import ChatAnthropic  # pylint: disable=import-error
+from langchain.chat_models import init_chat_model  # pylint: disable=import-error
 
 # pylint: disable=import-error
 from langchain_daytona_data_analysis import DaytonaDataAnalysisTool
@@ -16,7 +17,11 @@
 
 load_dotenv()
 
-model = ChatAnthropic(model_name="claude-sonnet-4-5-20250929", temperature=0, timeout=None, max_retries=2, stop=None)
+# "provider:model" string, e.g. "fireworks:accounts/fireworks/models/kimi-k2p6".
+# See the README for supported providers and their API keys.
+MODEL = os.environ.get("MODEL", "anthropic:claude-sonnet-4-5-20250929")
+
+model = init_chat_model(MODEL, temperature=0, max_retries=2)
 
 
 def process_data_analysis_result(result: ExecutionArtifacts):

python/langgraph/plan-and-execute-data-agent/.env.example

@@ -1,4 +1,9 @@
 # Get your API key from https://app.daytona.io/dashboard/keys
 DAYTONA_API_KEY=
+# Model in "provider:model" format; defaults to the Anthropic model below.
+# MODEL=anthropic:claude-opus-4-6
+# Set the API key matching your chosen provider:
 # Get your Anthropic API key from https://console.anthropic.com/
-ANTHROPIC_API_KEY=
+ANTHROPIC_API_KEY=
+# FIREWORKS_API_KEY=
+# OPENAI_API_KEY=

python/langgraph/plan-and-execute-data-agent/README.md

@@ -25,7 +25,11 @@ In this example, the agent profiles the maintenance health of the public [`langc
 ## Environment Variables
 
 - `DAYTONA_API_KEY`: Required for access to Daytona sandboxes. Get it from [Daytona Dashboard](https://app.daytona.io/dashboard/keys)
-- `ANTHROPIC_API_KEY`: Required for Claude model access. Get it from [Anthropic Console](https://console.anthropic.com/)
+- `MODEL` (optional): The model in `provider:model` format. Defaults to `anthropic:claude-opus-4-6`
+- The API key matching your chosen provider:
+  - `ANTHROPIC_API_KEY`: Required if using Anthropic models (default)
+  - `FIREWORKS_API_KEY`: Required if using Fireworks AI models
+  - `OPENAI_API_KEY`: Required if using OpenAI models
 
 Copy `.env.example` to `.env` and fill in your API keys before running.
 
@@ -41,7 +45,7 @@ Copy `.env.example` to `.env` and fill in your API keys before running.
 2. Install dependencies:
 
    ```bash
-   pip install -U langgraph langchain-core langchain-anthropic daytona pydantic python-dotenv
+   pip install -U langgraph langchain langchain-anthropic daytona pydantic python-dotenv
    ```
 
 3. Set your API keys in `.env`:
@@ -51,6 +55,17 @@ Copy `.env.example` to `.env` and fill in your API keys before running.
    ANTHROPIC_API_KEY=your_anthropic_api_key
    ```
 
+   The model is set via the `MODEL` environment variable in `provider:model` format and works with any [provider supported by LangChain](https://docs.langchain.com/oss/python/integrations/chat). Install the provider's integration package and set its API key. For example, to use [Fireworks AI](https://fireworks.ai/):
+
+   ```bash
+   pip install -U langchain-fireworks
+   ```
+
+   ```bash
+   MODEL=fireworks:accounts/fireworks/models/kimi-k2p6
+   FIREWORKS_API_KEY=your_fireworks_api_key
+   ```
+
 4. Run the example:
 
    ```bash

python/langgraph/plan-and-execute-data-agent/main.py

@@ -3,11 +3,13 @@
 
 """Plan-and-execute LangGraph agent that ETLs GitHub data into a Daytona-sandboxed SQLite."""
 
+import os
 import re
 from typing import TypedDict
 
 from dotenv import load_dotenv  # pylint: disable=import-error
-from langchain_anthropic import ChatAnthropic  # pylint: disable=import-error
+from langchain.chat_models import init_chat_model  # pylint: disable=import-error
+from langchain_core.language_models import BaseChatModel  # pylint: disable=import-error
 from langchain_core.messages import HumanMessage, SystemMessage  # pylint: disable=import-error
 from langgraph.graph import END, START, StateGraph  # pylint: disable=import-error
 from pydantic import BaseModel, Field  # pylint: disable=import-error
@@ -16,6 +18,10 @@
 
 load_dotenv()
 
+# "provider:model" string, e.g. "fireworks:accounts/fireworks/models/kimi-k2p6".
+# See the README for supported providers and their API keys.
+MODEL = os.environ.get("MODEL", "anthropic:claude-opus-4-6")
+
 USER_REQUEST = """Profile the maintenance health of the public GitHub repository `langchain-ai/langgraph`.
 
 Fetch the 100 most recently updated issues from
@@ -108,7 +114,7 @@ def extract_code(text: str) -> str:
     return match.group(1).strip() if match else text.strip()
 
 
-def build_graph(model: ChatAnthropic):
+def build_graph(model: BaseChatModel):
     """Wire the 6-node plan-and-execute state graph."""
     plan_llm = model.with_structured_output(Plan)
 
@@ -248,13 +254,7 @@ def cleanup(state: AgentState) -> dict:
 
 
 def main() -> None:
-    model = ChatAnthropic(
-        model_name="claude-opus-4-6",
-        temperature=0,
-        timeout=120,
-        max_retries=2,
-        stop=None,
-    )
+    model = init_chat_model(MODEL, temperature=0, timeout=120, max_retries=2)
     app = build_graph(model)
 
     print("=" * 60)

typescript/mastra/coding-agent/.env.example

@@ -4,13 +4,15 @@
 #   openai/gpt-4o-mini
 #   anthropic/claude-sonnet-4-5-20250929
 #   google/gemini-2.5-pro
+#   fireworks-ai/accounts/fireworks/models/kimi-k2p6
 #   groq/llama-3.3-70b-versatile
 MODEL=openai/gpt-4o-mini
 
 # API Keys (only needed for your chosen provider)
 OPENAI_API_KEY=
 ANTHROPIC_API_KEY=
 GOOGLE_GENERATIVE_AI_API_KEY=
+FIREWORKS_API_KEY=
 GROQ_API_KEY=
 CEREBRAS_API_KEY=
 MISTRAL_API_KEY=

typescript/mastra/coding-agent/README.md

@@ -59,6 +59,7 @@ This template supports any AI model provider through Mastra's model router. You
 - **OpenAI**: `openai/gpt-4o-mini`, `openai/gpt-4o`
 - **Anthropic**: `anthropic/claude-sonnet-4-5-20250929`, `anthropic/claude-haiku-4-5-20250929`
 - **Google**: `google/gemini-2.5-pro`, `google/gemini-2.0-flash-exp`
+- **Fireworks AI**: `fireworks-ai/accounts/fireworks/models/kimi-k2p6`, `fireworks-ai/accounts/fireworks/models/glm-5p2`
 - **Groq**: `groq/llama-3.3-70b-versatile`, `groq/llama-3.1-8b-instant`
 - **Cerebras**: `cerebras/llama-3.3-70b`
 - **Mistral**: `mistral/mistral-medium-2508`

typescript/vercel-ai-sdk/multi-language-benchmark-agent/README.md

@@ -135,12 +135,19 @@ Worth noticing in the transcript:
 - **Step 5 batches parallel tool calls.** Both `downloadFile` calls go out in the same turn and the Vercel AI SDK runs them concurrently, so the chart and the report ship to local disk at the same time rather than back-to-back.
 - **Steps 1, 2, 3 all use `runCode`** to execute Python, TypeScript, and Python (matplotlib) in turn. No `runCommand` was needed for the TypeScript step because `runCode` handles `ts-node` invocation internally. The exact step count and tool mix will vary run to run; some runs may use `runCommand` if the agent decides to `pip install` an extra package or inspect the filesystem before plotting.
 
-## Switching to OpenAI
+## Switching providers
 
-Install [`@ai-sdk/openai`](https://www.npmjs.com/package/@ai-sdk/openai) and replace the provider import in `src/index.ts`:
+The default is Anthropic Claude Sonnet 4.6, but any AI SDK provider that supports tool calling works. Install the provider package and replace the provider import in `src/index.ts`.
+
+> [!NOTE]
+> The provider package major version must match this guide's `ai` v6: use `@ai-sdk/openai@3` / `@ai-sdk/fireworks@2`. If you upgrade `ai` to v7, install the latest provider versions instead.
+
+### OpenAI
+
+Install [`@ai-sdk/openai`](https://www.npmjs.com/package/@ai-sdk/openai):
 
 ```bash
-npm install @ai-sdk/openai
+npm install @ai-sdk/openai@3
 ```
 
 ```typescript
@@ -149,4 +156,20 @@ import { openai } from '@ai-sdk/openai'
 const MODEL = openai('gpt-4o')
 ```
 
-Set `OPENAI_API_KEY` in `.env` instead of `ANTHROPIC_API_KEY`. The default is Anthropic Claude Sonnet 4.6, but any AI SDK provider that supports tool calling works.
+Set `OPENAI_API_KEY` in `.env` instead of `ANTHROPIC_API_KEY`.
+
+### Fireworks AI
+
+Install [`@ai-sdk/fireworks`](https://www.npmjs.com/package/@ai-sdk/fireworks):
+
+```bash
+npm install @ai-sdk/fireworks@2
+```
+
+```typescript
+import { fireworks } from '@ai-sdk/fireworks'
+// ...
+const MODEL = fireworks('accounts/fireworks/models/kimi-k2p6')
+```
+
+Set `FIREWORKS_API_KEY` in `.env` instead of `ANTHROPIC_API_KEY`.