{ "@context": "https://schema.org", "@id": "https://aggr.aymericbeaumet.com/items/philschmid-de/2026-01-22-gemini-interactions-api-quick-start/#webpage", "@type": [ "WebPage", "ArchiveComponent" ], "articleBody": "The [Interactions API](https://ai.google.dev/gemini-api/docs/interactions?ua=chat) is a unified interface for building with Gemini models and agents. It simplifies the development of agentic applications by handling server-side state management, tool orchestration, and long-running tasks.\n\nWith a single endpoint, you can:\n\n- Interact with Gemini models for text, image, and audio generation\n- Build multi-turn conversations without managing history client-side\n- Call custom functions and built-in tools like Google Search\n- Run specialized agents like Deep Research for complex tasks\n\n[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/google-gemini/cookbook/blob/main/quickstarts/Get_started_interactions_api.ipynb)\n\n## Prerequisites\n\nGet a [Gemini API key](https://aistudio.google.com/apikey) and install the Google GenAI SDK:\n\nBash\n\n```bash\npip install google-genai\n```\n\nSet your API key:\n\nBash\n\n```bash\nexport GEMINI_API_KEY=\"your-api-key\"\n```\n\n## Create an interaction\n\nAt its simplest, the Interactions API works like a standard chat completion. You provide a model and an input string. By default, interactions are stored (`store=True`), allowing you to reference them later.\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n system_instruction=\"You are a helpful assistant.\",\n input=\"Explain quantum entanglement in one sentence.\"\n)\n \nprint(interaction.outputs[-1].text)\n# Output: Quantum entanglement is a phenomenon where particles become linked...\n```\n\n## Stateful Multi-turn Conversations\n\nOne of the most powerful features of this API is **server-side state management**. You do not need to append messages to a list and send the full history back to the server every time.\n\nUse `previous_interaction_id` to continue a conversation.\n\nNote: Only conversation history is preserved. Parameters like `tools` or `generation_config` are interaction-scoped and must be re-declared if needed.\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \nturn_1 = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"My name is Alice and I am a software engineer.\"\n)\nprint(f\"Turn 1 ID: {turn_1.id}\")\n \nturn_2 = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What is my job?\",\n previous_interaction_id=turn_1.id\n)\n \nprint(turn_2.outputs[-1].text)\n# Output: You are a software engineer.\n```\n\nFor client-managed history, see [Stateless conversations](https://www.philschmid.de/gemini-api/docs/interactions#stateless-conversation).\n\n### Forking Conversations\n\nBecause state is managed by ID, you can \"fork\" a conversation by referencing an older interaction ID with a different prompt.\n\nPython\n\n```python\n# Branch off from Turn 1 with a different topic\nturn_2_fork = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What is my name?\",\n previous_interaction_id=turn_1.id\n)\n \nprint(turn_2_fork.outputs[-1].text)\n# Output: Your name is Alice.\n```\n\n## Multimodal Interactions\n\nGemini models natively understand and generate multiple content types. You can pass text, images, audio, or PDF documents in a single interaction. This example uses a remote image URL.\n\n### Multimodal understanding\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=[\n {\"type\": \"text\", \"text\": \"Generate a recipe for the shown scones.\"},\n {\n \"type\": \"image\", \n \"uri\": \"https://storage.googleapis.com/generativeai-downloads/images/scones.jpg\"\n }\n ],\n)\n \nprint(interaction.outputs[-1].text)\n```\n\nFor audio, video, and document (PDF) understanding, see [Multimodal understanding](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#understanding).\n\n### Multimodal Generation\n\nPython\n\n```python\nimport base64\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-pro-image-preview\",\n input=\"Generate an image of a futuristic city at sunset.\"\n)\n \nfor output in interaction.outputs:\n if output.type == \"image\":\n with open(\"city.png\", \"wb\") as f:\n f.write(base64.b64decode(output.data))\n```\n\nFor audio generation see [Multimodal generations](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#generation).\n\n## Tool use\n\nTools extend the model's capabilities by letting it call external functions or services. The API includes ready-to-use tools like Google Search, and lets you define custom tools as JSON schemas. The model decides when to call them based on the conversation:\n\n### Built-in tools\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"Who won the 2024 Nobel Prize in Physics?\",\n tools=[{\"type\": \"google_search\"}]\n)\n \ntext_output = next((o for o in interaction.outputs if o.type == \"text\"), None)\nif text_output:\n print(text_output.text)\n```\n\nOther built-in tools include [Code Execution](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling) and [Computer Use](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling).\n\n### Function calling\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \n# Define a tool\nweather_tool = {\n \"type\": \"function\",\n \"name\": \"get_weather\",\n \"description\": \"Get current weather for a location\",\n \"parameters\": {\n \"type\": \"object\",\n \"properties\": {\n \"location\": {\"type\": \"string\", \"description\": \"City name\"}\n },\n \"required\": [\"location\"]\n }\n}\n \n# Send request with tool\ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What's the weather in Tokyo?\",\n tools=[weather_tool]\n)\n \n# Handle tool call\nfor output in interaction.outputs:\n if output.type == \"function_call\":\n # Execute your function (mocked here)\n # result = get_weather(output.arguments\n result = {\"temperature\": \"22°C\", \"condition\": \"sunny\"}\n \n # Return result to model\n interaction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n previous_interaction_id=interaction.id,\n input={\n \"type\": \"function_result\",\n \"name\": output.name,\n \"call_id\": output.id,\n \"result\": result\n }\n )\n print(interaction.outputs[-1].text)\n```\n\nFor code execution, URL context, and MCP servers, see [Agentic capabilities](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling).\n\n## Agents & Long-Running Tasks\n\nBeyond models, the Interactions API provides access to specialized agents. Deep Research executes multi-step research tasks, synthesizing information from multiple sources into comprehensive reports.\n\nAgents run asynchronously with `background=True`. Poll the interaction status to retrieve results:\n\nPython\n\n```python\nimport time\nfrom google import genai\n \nclient = genai.Client()\n \nagent_interaction = client.interactions.create(\n agent=\"deep-research-pro-preview-12-2025\", # Note: use 'agent', not 'model'\n input=\"Research the history of the Google TPUs with a focus on 2025 specs.\",\n background=True\n)\n \n \n# Poll for completion\nwhile True:\n status_check = client.interactions.get(agent_interaction.id)\n print(f\"Status: {status_check.status}\")\n \n if status_check.status == \"completed\":\n print(\"\\n--- Final Report ---\\n\")\n print(status_check.outputs[-1].text)\n break\n elif status_check.status in [\"failed\", \"cancelled\"]:\n print(\"Agent failed.\")\n break\n \n time.sleep(10)\n```\n\nFor more details, see [Deep Research](https://ai.google.dev/gemini-api/docs/deep-research).\n\n## Next Steps\n\nThe Interactions API supports many more complex workflows. Check the [API Reference](https://ai.google.dev/api/interactions-api) for details on:\n\n- **[Structured Outputs](https://ai.google.dev/gemini-api/docs/structured-output):** Force the model to return valid JSON matching a specific schema.\n- **[Streaming](https://ai.google.dev/gemini-api/docs/interactions#streaming):** Stream token responses for real-time applications.\n- **[Thinking Models](https://ai.google.dev/api/interactions-api#thinking):** Configure `thinking_level` for Gemini 2.5 and 3.0 models to handle complex reasoning.\n- **[Remote MCP](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#remote-mcp-model-context-protocol):** Connect Gemini to your own private MCP servers.\n- **[Combining tools and structured outputs](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#combining-tools-and-structured-output):** Combine tools and structured outputs to create more complex workflows.\n- **[File Uploads](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#working-with-files):** Upload files to the model for processing.\n- **[Data Model](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#data-model):** high level overview of the main inputs and outputs of the API.\n\n## Feedback\n\n**This API is in Beta, and we want your feedback!** We're actively listening to developers to shape the future of this API. What features would help your agent workflows? What pain points are you experiencing? Please let me know on [Twitter](https://twitter.com/_philschmid) or [LinkedIn](https://www.linkedin.com/in/philipp-schmid-a6a2bb196/).\n", "authors": [], "captured_at": "2026-09-27T19:29:15.293681927+00:00", "category": "programming", "content_html": "

The Interactions API is a unified interface for building with Gemini models and agents. It simplifies the development of agentic applications by handling server-side state management, tool orchestration, and long-running tasks.

\n

With a single endpoint, you can:

\n\n

\"Open

\n

Prerequisites

\n

Get a Gemini API key and install the Google GenAI SDK:

\n

Bash

\n
pip install google-genai\n
\n

Set your API key:

\n

Bash

\n
export GEMINI_API_KEY="your-api-key"\n
\n

Create an interaction

\n

At its simplest, the Interactions API works like a standard chat completion. You provide a model and an input string. By default, interactions are stored (store=True), allowing you to reference them later.

\n

Python

\n
from google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n    model="gemini-3-flash-preview",\n    system_instruction="You are a helpful assistant.",\n    input="Explain quantum entanglement in one sentence."\n)\n \nprint(interaction.outputs[-1].text)\n# Output: Quantum entanglement is a phenomenon where particles become linked...\n
\n

Stateful Multi-turn Conversations

\n

One of the most powerful features of this API is server-side state management. You do not need to append messages to a list and send the full history back to the server every time.

\n

Use previous_interaction_id to continue a conversation.

\n

Note: Only conversation history is preserved. Parameters like tools or generation_config are interaction-scoped and must be re-declared if needed.

\n

Python

\n
from google import genai\n \nclient = genai.Client()\n \nturn_1 = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input="My name is Alice and I am a software engineer."\n)\nprint(f"Turn 1 ID: {turn_1.id}")\n \nturn_2 = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input="What is my job?",\n    previous_interaction_id=turn_1.id\n)\n \nprint(turn_2.outputs[-1].text)\n# Output: You are a software engineer.\n
\n

For client-managed history, see Stateless conversations.

\n

Forking Conversations

\n

Because state is managed by ID, you can "fork" a conversation by referencing an older interaction ID with a different prompt.

\n

Python

\n
# Branch off from Turn 1 with a different topic\nturn_2_fork = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input="What is my name?",\n    previous_interaction_id=turn_1.id\n)\n \nprint(turn_2_fork.outputs[-1].text)\n# Output: Your name is Alice.\n
\n

Multimodal Interactions

\n

Gemini models natively understand and generate multiple content types. You can pass text, images, audio, or PDF documents in a single interaction. This example uses a remote image URL.

\n

Multimodal understanding

\n

Python

\n
from google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input=[\n        {"type": "text", "text": "Generate a recipe for the shown scones."},\n        {\n            "type": "image", \n            "uri": "https://storage.googleapis.com/generativeai-downloads/images/scones.jpg"\n        }\n    ],\n)\n \nprint(interaction.outputs[-1].text)\n
\n

For audio, video, and document (PDF) understanding, see Multimodal understanding.

\n

Multimodal Generation

\n

Python

\n
import base64\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n    model="gemini-3-pro-image-preview",\n    input="Generate an image of a futuristic city at sunset."\n)\n \nfor output in interaction.outputs:\n    if output.type == "image":\n        with open("city.png", "wb") as f:\n            f.write(base64.b64decode(output.data))\n
\n

For audio generation see Multimodal generations.

\n

Tool use

\n

Tools extend the model's capabilities by letting it call external functions or services. The API includes ready-to-use tools like Google Search, and lets you define custom tools as JSON schemas. The model decides when to call them based on the conversation:

\n

Built-in tools

\n

Python

\n
from google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input="Who won the 2024 Nobel Prize in Physics?",\n    tools=[{"type": "google_search"}]\n)\n \ntext_output = next((o for o in interaction.outputs if o.type == "text"), None)\nif text_output:\n    print(text_output.text)\n
\n

Other built-in tools include Code Execution and Computer Use.

\n

Function calling

\n

Python

\n
from google import genai\n \nclient = genai.Client()\n \n# Define a tool\nweather_tool = {\n    "type": "function",\n    "name": "get_weather",\n    "description": "Get current weather for a location",\n    "parameters": {\n        "type": "object",\n        "properties": {\n            "location": {"type": "string", "description": "City name"}\n        },\n        "required": ["location"]\n    }\n}\n \n# Send request with tool\ninteraction = client.interactions.create(\n    model="gemini-3-flash-preview",\n    input="What's the weather in Tokyo?",\n    tools=[weather_tool]\n)\n \n# Handle tool call\nfor output in interaction.outputs:\n    if output.type == "function_call":\n        # Execute your function (mocked here)\n        # result = get_weather(output.arguments\n        result = {"temperature": "22°C", "condition": "sunny"}\n \n        # Return result to model\n        interaction = client.interactions.create(\n            model="gemini-3-flash-preview",\n            previous_interaction_id=interaction.id,\n            input={\n                "type": "function_result",\n                "name": output.name,\n                "call_id": output.id,\n                "result": result\n            }\n        )\n        print(interaction.outputs[-1].text)\n
\n

For code execution, URL context, and MCP servers, see Agentic capabilities.

\n

Agents & Long-Running Tasks

\n

Beyond models, the Interactions API provides access to specialized agents. Deep Research executes multi-step research tasks, synthesizing information from multiple sources into comprehensive reports.

\n

Agents run asynchronously with background=True. Poll the interaction status to retrieve results:

\n

Python

\n
import time\nfrom google import genai\n \nclient = genai.Client()\n \nagent_interaction = client.interactions.create(\n    agent="deep-research-pro-preview-12-2025", # Note: use 'agent', not 'model'\n    input="Research the history of the Google TPUs with a focus on 2025 specs.",\n    background=True\n)\n \n \n# Poll for completion\nwhile True:\n    status_check = client.interactions.get(agent_interaction.id)\n    print(f"Status: {status_check.status}")\n    \n    if status_check.status == "completed":\n        print("\\n--- Final Report ---\\n")\n        print(status_check.outputs[-1].text)\n        break\n    elif status_check.status in ["failed", "cancelled"]:\n        print("Agent failed.")\n        break\n        \n    time.sleep(10)\n
\n

For more details, see Deep Research.

\n

Next Steps

\n

The Interactions API supports many more complex workflows. Check the API Reference for details on:

\n\n

Feedback

\n

This API is in Beta, and we want your feedback! We're actively listening to developers to shape the future of this API. What features would help your agent workflows? What pain points are you experiencing? Please let me know on Twitter or LinkedIn.

\n", "content_markdown": "The [Interactions API](https://ai.google.dev/gemini-api/docs/interactions?ua=chat) is a unified interface for building with Gemini models and agents. It simplifies the development of agentic applications by handling server-side state management, tool orchestration, and long-running tasks.\n\nWith a single endpoint, you can:\n\n- Interact with Gemini models for text, image, and audio generation\n- Build multi-turn conversations without managing history client-side\n- Call custom functions and built-in tools like Google Search\n- Run specialized agents like Deep Research for complex tasks\n\n[![Open In Colab](https://colab.research.google.com/assets/colab-badge.svg)](https://colab.research.google.com/github/google-gemini/cookbook/blob/main/quickstarts/Get_started_interactions_api.ipynb)\n\n## Prerequisites\n\nGet a [Gemini API key](https://aistudio.google.com/apikey) and install the Google GenAI SDK:\n\nBash\n\n```bash\npip install google-genai\n```\n\nSet your API key:\n\nBash\n\n```bash\nexport GEMINI_API_KEY=\"your-api-key\"\n```\n\n## Create an interaction\n\nAt its simplest, the Interactions API works like a standard chat completion. You provide a model and an input string. By default, interactions are stored (`store=True`), allowing you to reference them later.\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n system_instruction=\"You are a helpful assistant.\",\n input=\"Explain quantum entanglement in one sentence.\"\n)\n \nprint(interaction.outputs[-1].text)\n# Output: Quantum entanglement is a phenomenon where particles become linked...\n```\n\n## Stateful Multi-turn Conversations\n\nOne of the most powerful features of this API is **server-side state management**. You do not need to append messages to a list and send the full history back to the server every time.\n\nUse `previous_interaction_id` to continue a conversation.\n\nNote: Only conversation history is preserved. Parameters like `tools` or `generation_config` are interaction-scoped and must be re-declared if needed.\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \nturn_1 = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"My name is Alice and I am a software engineer.\"\n)\nprint(f\"Turn 1 ID: {turn_1.id}\")\n \nturn_2 = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What is my job?\",\n previous_interaction_id=turn_1.id\n)\n \nprint(turn_2.outputs[-1].text)\n# Output: You are a software engineer.\n```\n\nFor client-managed history, see [Stateless conversations](https://www.philschmid.de/gemini-api/docs/interactions#stateless-conversation).\n\n### Forking Conversations\n\nBecause state is managed by ID, you can \"fork\" a conversation by referencing an older interaction ID with a different prompt.\n\nPython\n\n```python\n# Branch off from Turn 1 with a different topic\nturn_2_fork = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What is my name?\",\n previous_interaction_id=turn_1.id\n)\n \nprint(turn_2_fork.outputs[-1].text)\n# Output: Your name is Alice.\n```\n\n## Multimodal Interactions\n\nGemini models natively understand and generate multiple content types. You can pass text, images, audio, or PDF documents in a single interaction. This example uses a remote image URL.\n\n### Multimodal understanding\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=[\n {\"type\": \"text\", \"text\": \"Generate a recipe for the shown scones.\"},\n {\n \"type\": \"image\", \n \"uri\": \"https://storage.googleapis.com/generativeai-downloads/images/scones.jpg\"\n }\n ],\n)\n \nprint(interaction.outputs[-1].text)\n```\n\nFor audio, video, and document (PDF) understanding, see [Multimodal understanding](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#understanding).\n\n### Multimodal Generation\n\nPython\n\n```python\nimport base64\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-pro-image-preview\",\n input=\"Generate an image of a futuristic city at sunset.\"\n)\n \nfor output in interaction.outputs:\n if output.type == \"image\":\n with open(\"city.png\", \"wb\") as f:\n f.write(base64.b64decode(output.data))\n```\n\nFor audio generation see [Multimodal generations](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#generation).\n\n## Tool use\n\nTools extend the model's capabilities by letting it call external functions or services. The API includes ready-to-use tools like Google Search, and lets you define custom tools as JSON schemas. The model decides when to call them based on the conversation:\n\n### Built-in tools\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"Who won the 2024 Nobel Prize in Physics?\",\n tools=[{\"type\": \"google_search\"}]\n)\n \ntext_output = next((o for o in interaction.outputs if o.type == \"text\"), None)\nif text_output:\n print(text_output.text)\n```\n\nOther built-in tools include [Code Execution](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling) and [Computer Use](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling).\n\n### Function calling\n\nPython\n\n```python\nfrom google import genai\n \nclient = genai.Client()\n \n# Define a tool\nweather_tool = {\n \"type\": \"function\",\n \"name\": \"get_weather\",\n \"description\": \"Get current weather for a location\",\n \"parameters\": {\n \"type\": \"object\",\n \"properties\": {\n \"location\": {\"type\": \"string\", \"description\": \"City name\"}\n },\n \"required\": [\"location\"]\n }\n}\n \n# Send request with tool\ninteraction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n input=\"What's the weather in Tokyo?\",\n tools=[weather_tool]\n)\n \n# Handle tool call\nfor output in interaction.outputs:\n if output.type == \"function_call\":\n # Execute your function (mocked here)\n # result = get_weather(output.arguments\n result = {\"temperature\": \"22°C\", \"condition\": \"sunny\"}\n \n # Return result to model\n interaction = client.interactions.create(\n model=\"gemini-3-flash-preview\",\n previous_interaction_id=interaction.id,\n input={\n \"type\": \"function_result\",\n \"name\": output.name,\n \"call_id\": output.id,\n \"result\": result\n }\n )\n print(interaction.outputs[-1].text)\n```\n\nFor code execution, URL context, and MCP servers, see [Agentic capabilities](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#tools-and-function-calling).\n\n## Agents & Long-Running Tasks\n\nBeyond models, the Interactions API provides access to specialized agents. Deep Research executes multi-step research tasks, synthesizing information from multiple sources into comprehensive reports.\n\nAgents run asynchronously with `background=True`. Poll the interaction status to retrieve results:\n\nPython\n\n```python\nimport time\nfrom google import genai\n \nclient = genai.Client()\n \nagent_interaction = client.interactions.create(\n agent=\"deep-research-pro-preview-12-2025\", # Note: use 'agent', not 'model'\n input=\"Research the history of the Google TPUs with a focus on 2025 specs.\",\n background=True\n)\n \n \n# Poll for completion\nwhile True:\n status_check = client.interactions.get(agent_interaction.id)\n print(f\"Status: {status_check.status}\")\n \n if status_check.status == \"completed\":\n print(\"\\n--- Final Report ---\\n\")\n print(status_check.outputs[-1].text)\n break\n elif status_check.status in [\"failed\", \"cancelled\"]:\n print(\"Agent failed.\")\n break\n \n time.sleep(10)\n```\n\nFor more details, see [Deep Research](https://ai.google.dev/gemini-api/docs/deep-research).\n\n## Next Steps\n\nThe Interactions API supports many more complex workflows. Check the [API Reference](https://ai.google.dev/api/interactions-api) for details on:\n\n- **[Structured Outputs](https://ai.google.dev/gemini-api/docs/structured-output):** Force the model to return valid JSON matching a specific schema.\n- **[Streaming](https://ai.google.dev/gemini-api/docs/interactions#streaming):** Stream token responses for real-time applications.\n- **[Thinking Models](https://ai.google.dev/api/interactions-api#thinking):** Configure `thinking_level` for Gemini 2.5 and 3.0 models to handle complex reasoning.\n- **[Remote MCP](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#remote-mcp-model-context-protocol):** Connect Gemini to your own private MCP servers.\n- **[Combining tools and structured outputs](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#combining-tools-and-structured-output):** Combine tools and structured outputs to create more complex workflows.\n- **[File Uploads](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#working-with-files):** Upload files to the model for processing.\n- **[Data Model](https://ai.google.dev/gemini-api/docs/interactions?ua=chat#data-model):** high level overview of the main inputs and outputs of the API.\n\n## Feedback\n\n**This API is in Beta, and we want your feedback!** We're actively listening to developers to shape the future of this API. What features would help your agent workflows? What pain points are you experiencing? Please let me know on [Twitter](https://twitter.com/_philschmid) or [LinkedIn](https://www.linkedin.com/in/philipp-schmid-a6a2bb196/).\n", "dateCreated": "2026-09-27T19:29:15.293681927+00:00", "date_created": "2026-01-22T00:00:00+00:00", "date_modified": "2026-01-22T00:00:00+00:00", "date_published": "2026-01-22T00:00:00+00:00", "description": "The Interactions API is a unified interface for building with Gemini models and agents. It simplifies the development of agentic applications by handling server-side state management, tool orchestration, and long-running tasks.", "external_url": "https://www.philschmid.de/interactions-api-quickstart", "git": { "commit": "ebb0941c5c06a2ab2377342d607535262a3e969e", "path": "items/philschmid-de/2026/01/2026-01-22-gemini-interactions-api-quick-start.md", "permalink": "https://github.com/aymericbeaumet/aggr-instance/blob/ebb0941c5c06a2ab2377342d607535262a3e969e/items/philschmid-de/2026/01/2026-01-22-gemini-interactions-api-quick-start.md" }, "id": "urn:aggr:item:37f41fd61d4d955b207530878cff6ee4e110b062", "isBasedOn": { "@id": "https://www.philschmid.de/interactions-api-quickstart" }, "mainEntity": { "@id": "https://www.philschmid.de/interactions-api-quickstart", "@type": "CreativeWork", "archivedAt": { "@id": "https://aggr.aymericbeaumet.com/items/philschmid-de/2026-01-22-gemini-interactions-api-quick-start/#webpage" }, "dateModified": "2026-01-22T00:00:00+00:00", "datePublished": "2026-01-22T00:00:00+00:00", "headline": "Gemini Interactions API Quick Start", "timeRequired": "PT5M", "url": "https://www.philschmid.de/interactions-api-quickstart", "wordCount": 984 }, "name": "Archived snapshot: Gemini Interactions API Quick Start", "preview": null, "reading_minutes": 5, "replicated_at": null, "source": "philschmid-de", "source_name": "philschmid.de - RSS feed", "summary": "The Interactions API is a unified interface for building with Gemini models and agents. It simplifies the development of agentic applications by handling server-side state management, tool orchestration, and long-running tasks.", "tags": [], "temporalCoverage": "2026-09-27T19:29:15.293681927+00:00", "title": "Gemini Interactions API Quick Start", "url": "https://aggr.aymericbeaumet.com/items/philschmid-de/2026-01-22-gemini-interactions-api-quick-start/", "word_count": 984 }