references/migration.md
# Migration Reference
How to migrate existing Gemini API code to the Interactions API and/or upgrade between model generations. Covers the agent workflow for performing migrations safely.
For detailed before/after code examples across all feature areas (text generation, multi-turn, streaming, function calling, structured output, grounding, multimodal), fetch the full migration guide: https://aistudio.google.com/docs/migrate-to-interactions.md
## Confirm the Migration Scope
**Before any edits, confirm the scope.** If the user's request does not explicitly name a single file, a specific directory, or an explicit file list, ask first and do not start editing.
Even imperative requests like "migrate my code", "upgrade to gemini 3", "migrate my app to gemini", or "switch to the Interactions API" leave the scope ambiguous. Ask:
> Before I start editing, can you confirm the scope?
> 1. Entire project
> 2. Specific subdirectory (e.g. `src/`, `api/`)
> 3. Specific file or list of files
**Sizing the scope (large repos).** Before asking, get a per-directory count:
```sh
rg -l "generate_content|generateContent|gemini-1\.5|gemini-2\.0|gemini-2\.5|gemini-3\.1-pro|gemini-3\.1-flash-lite|gemini-3\.1-flash-tts|gemini-3\.5|gemini-3\.6|gemini-3\.7|gemini-3-flash|gemini-omni-1\.1|gemini-omni-flash-preview|thinking_budget|temperature" --type-not md | cut -d/ -f1 | sort | uniq -c | sort -rn
```
Present the breakdown in your question (e.g. *"Found 42 references across 3 directories: src/ (28), tests/ (10), scripts/ (4). Which to migrate?"*).
**Proceed without asking** only when the scope is already unambiguous, the user named an exact file ("migrate `app.py`"), pointed at a directory ("migrate everything under `src/`"), or already confirmed scope in an earlier turn.
## API Migration: `generateContent` → `Interactions`
The core changes when migrating from `generateContent` to the Interactions API:
| What | `generateContent` | Interactions API |
|------|----------------|-----------------|
| **SDK method** | `client.models.generate_content()` | `client.interactions.create()` |
| **Response text** | `response.text` | `interaction.steps[-1].content[0].text` |
| **Multi-turn** | Manual history array or `client.chats.create()` | `previous_interaction_id=interaction.id` |
| **Streaming** | `generate_content_stream()` / `:streamGenerateContent` | `stream=True` + `step.delta` events |
| **Structured output** | `config.response_format` inside `GenerateContentConfig` | Top-level `response_format` array |
| **Function calling** | `candidates[0].content.parts[0].function_call` | `function_call` step in `interaction.steps` |
| **Search grounding** | `groundingMetadata` on candidates | `google_search_call`/`google_search_result` steps + inline `annotations` |
| **Config/types** | `types.GenerateContentConfig(...)`, `types.Tool(...)`, `types.Content(...)`, `types.Part.*` | Not used. Interactions API uses plain Python dicts and direct params. Check the feature docs for exact format. |
| **REST endpoint** | `POST /v1beta/models/{model}:generateContent` | `POST /v1beta/interactions` |
| **SDK package** | `google-genai` ≥ 1.x or legacy `google-generativeai` | `google-genai` ≥ 2.30.1 (Python) / `@google/genai` ≥ 2.29.1 (JS) |
For full before/after code examples, fetch the [Migration Guide](https://aistudio.google.com/docs/migrate-to-interactions.md) or read the Interactions API documentation pages for each feature.
## Model Migration
### Deprecated Models
| Model | Status | Drop-in Replacement |
|-------|--------|-------------------|
| `gemini-2.0-flash` | Deprecated | `gemini-3.8-flash` |
| `gemini-2.0-flash-lite` | Deprecated | `gemini-3.5-flash-lite` |
| `gemini-1.5-pro` | Deprecated | `gemini-3.8-flash` |
| `gemini-1.5-flash` | Deprecated | `gemini-3.8-flash` |
| `gemini-3.1-flash-image` | Deprecated (shutdown October 29, 2026) | `gemini-nano-banana-2.1` |
### Active Legacy Models (migration recommended)
| Current Model | Recommended Target | Why |
|--------------|-------------------|-----|
| `gemini-3.7-flash`, `gemini-3.6-flash`, `gemini-3.5-flash`, or `gemini-3-flash-preview` | `gemini-3.8-flash` | Latest Flash: stronger agentic/multimodal performance, reduced token usage/loop spiraling |
| `gemini-2.5-flash` | `gemini-3.8-flash` or `gemini-3.5-flash-lite` | Latest Flash with Interactions API support, or latest Flash-Lite for cheaper/simpler tasks. |
| `gemini-2.5-flash-lite` or `gemini-3.1-flash-lite` | `gemini-3.5-flash-lite` | Latest Flash-lite with Interactions API support |
| `gemini-3.1-pro-preview` or `gemini-2.5-pro` | `gemini-4-argon` | Most powerful model with 1M context, complex reasoning, coding, and research |
| `gemini-3.1-flash-tts-preview` or `gemini-2.5-*-tts` | `gemini-3.8-flash-tts` or `gemini-3.8-flash-lite-tts` | Latest TTS models with Voice Design, Voice Replication, structured `speech_metadata`, and default WAV (`audio/wav`) output |
| `gemini-3.1-flash-image` or older image generation models | `gemini-nano-banana-2.1` | Latest Nano Banana image generation model: 131k context, multi-image fusion (up to 14 reference images), enhanced visual quality, and text rendering |
| `gemini-omni-1.1-flash` or `gemini-omni-flash-preview` | `gemini-omni-1.2-flash` | Latest Omni Flash video model: `3s`–`15s` generation and extension (`10s` default; `3s`–`10s` two-frame keyframe interpolation), improved motion/character consistency, and natural-language media binding |
| Legacy audio understanding / ASR | `gemini-3.5-transcribe` | Dedicated speech-to-text with auto language detection, diarization, word timestamps, and smart formatting |
| Legacy Live API (`gemini-3.1-flash-live-preview`, `gemini-2.5-flash-native-audio-*`, `gemini-2.0-flash-live-001`) | `gemini-3.8-live` or `gemini-3.8-live-extended-thinking` | Install `google-gemini/gemini-live-api-dev` skill and see its `references/migration.md` for Live API protocol changes |
### Managed Agents
| Agent | Status | Drop-in Replacement | Notes |
|-------|--------|---------------------|-------|
| `antigravity-preview-05-2026` | Deprecated (October 5, 2026) | `antigravity-preview-09-2026` | Requests redirect to `antigravity-preview-09-2026` after October 5, 2026. |
| `deep-research-pro-preview-12-2025` | Deprecated (shutdown October 23, 2026) | `deep-research-preview-04-2026` or `deep-research-max-preview-04-2026` | Update `agent` parameter in `interactions.create`. |
> **Note:** Within the Interactions API, model upgrades are generally drop-in — change the model string and verify. The breaking changes are at the **API level** (generateContent → Interactions) and parameter deprecations (`temperature`, `top_p`, `top_k`).
## Migration Checklist
Every item is tagged: **`[BLOCKS]`** items cause errors or broken behavior if missed. **`[TUNE]`** items are quality/performance adjustments.
### API Migration (generateContent → Interactions)
- [ ] Updated SDK: `google-genai` ≥ 2.30.1 (Python) / `@google/genai` ≥ 2.29.1 (JS)
- [ ] Replaced `client.models.generate_content()` → `client.interactions.create()`
- [ ] Replaced `response.text` → `interaction.output_text`
- [ ] Replaced `response.candidates[0].content.parts` → iterate `interaction.steps`
- [ ] Replaced `client.chats.create()` / manual history → `previous_interaction_id`
- [ ] Removed all `types.*` wrappers (`GenerateContentConfig`, `Tool`, `Content`, `Part`) — Interactions API uses plain dicts. Check feature docs for exact format.
- [ ] Moved `response_format` from `GenerateContentConfig` to top-level parameter
- [ ] Replaced `generate_content_stream()` → `stream=True` + step-based event handling
- [ ] Updated function calling: candidates-based → step-based tool lifecycle
- [ ] REST: Changed endpoint to `/v1beta/interactions`
- [ ] REST: Add `Api-Revision: 2026-05-20` header (SDK ≥ 2.0.0 sets it automatically)
- [ ] Replaced `google-generativeai` (Python) → `google-genai` ≥ 2.30.1
- [ ] Replaced `@google/generative-ai` (JS) → `@google/genai` ≥ 2.29.1
- [ ] Updated all import statements to match new package names
### Model String Updates
- [ ] Replaced `gemini-2.0-*` model strings with current equivalents
- [ ] Replaced `gemini-1.5-*` model strings with current equivalents
- [ ] Consider upgrading `gemini-3.7-flash` → `gemini-3.8-flash`
- [ ] Consider upgrading `gemini-3.6-flash` → `gemini-3.8-flash`
- [ ] Consider upgrading `gemini-3.5-flash` → `gemini-3.8-flash`
- [ ] Consider upgrading `gemini-3-flash-preview` → `gemini-3.8-flash`
- [ ] Consider upgrading `gemini-2.5-flash` → `gemini-3.8-flash`
- [ ] Consider upgrading `gemini-3.1-flash-lite` → `gemini-3.5-flash-lite`
- [ ] Consider upgrading `gemini-2.5-flash-lite` → `gemini-3.5-flash-lite` or `gemini-3.1-flash-lite`
- [ ] Consider upgrading `gemini-3.1-pro-preview` → `gemini-4-argon`
- [ ] Consider upgrading `gemini-2.5-pro` → `gemini-4-argon`
- [ ] Consider upgrading `gemini-3.1-flash-tts-preview` → `gemini-3.8-flash-tts` or `gemini-3.8-flash-lite-tts`
- [ ] Consider upgrading `gemini-3.1-flash-image` → `gemini-nano-banana-2.1`
- [ ] Consider upgrading `gemini-omni-1.1-flash` → `gemini-omni-1.2-flash`
### Agent Updates
- [ ] Upgrade `antigravity-preview-05-2026` → `antigravity-preview-09-2026` (deprecated October 5, 2026; redirected afterwards)
- [ ] Upgrade `deep-research-pro-preview-12-2025` → `deep-research-preview-04-2026` or `deep-research-max-preview-04-2026` (deprecated, shutdown October 23, 2026)
### Migrate to Gemini 4 Argon
Use this checklist when migrating to `gemini-4-argon`. For full details on supported APIs, context window sizes, parameter changes, and continuation patterns, fetch [What's new in Gemini 4 Argon](https://aistudio.google.com/docs/latest-model.md) and the [Long Decode guide](https://aistudio.google.com/docs/long-decode.md).
- [ ] **Check for unsupported APIs before migrating**: Look up unsupported APIs and features in [What's new in Gemini 4 Argon](https://aistudio.google.com/docs/latest-model.md) (e.g., explicit context caching), ensure they are not used, and flag any unsupported usage to the user before migrating
- [ ] Updated SDK: `google-genai` ≥ 2.30.1 (Python) / `@google/genai` ≥ 2.29.1 (JS)
- [ ] Updated model name to `gemini-4-argon`
- [ ] Removed `temperature`, `top_p`, `top_k` from config
- [ ] Replaced `thinking_budget` with `thinking_level` (`low`, `medium`, `high`)
- [ ] Reviewed `max_output_tokens`: raise any limit below `32,768` to at least `32,768` (default cap is `65,536` tokens without long decode)
- [ ] **Decide whether long decode is needed**: Refer to the [Long Decode guide](https://aistudio.google.com/docs/long-decode.md) (input, thinking, and output share a single 1M token window). Only if needed for the workload:
- [ ] Set `max_output_tokens` explicitly above `65,536` (up to `1,048,576` minus input tokens)
- [ ] Enable [long decode](https://aistudio.google.com/docs/long-decode.md) via [background execution](https://aistudio.google.com/docs/background-execution.md) (`background=True`), poll while `interaction.status in ("in_progress", "queued")`, and join text across all `model_output` steps in `interaction.steps`
- [ ] Do not use `total_input_tokens` for context compaction, as it sums input tokens across all continuation hops
- [ ] Increase HTTP client read timeouts to at least 15–20 minutes
### Migrate to Gemini 3.8 Flash or Gemini 3.5 Flash-Lite
Use this checklist if the user requests to migrate to Gemini 3.8 Flash or Gemini 3.5 Flash-Lite. For full documentation of the changes, fetch the [What's new in Gemini 3.8 Flash guide](https://aistudio.google.com/docs/whats-new-gemini-3.8.md) and look for the migration section.
- [ ] Updated model name to `gemini-3.8-flash` or `gemini-3.5-flash-lite` (depending on user request)
- [ ] Removed `temperature`, `top_p`, `top_k` from config
- [ ] Replaced `thinking_budget` with `thinking_level` (`minimal`, `low`, `medium`, `high`)
### Migrate to Gemini 3.8 TTS (`gemini-3.8-flash-tts` or `gemini-3.8-flash-lite-tts`)
For full migration details, output formats, and prompting guide, fetch the [Speech Generation guide](https://aistudio.google.com/docs/speech-generation.md) (see `#migration` and `#prompting-guide`), [Voice Design](https://aistudio.google.com/docs/voice-design.md), and [Voice Replication](https://aistudio.google.com/docs/voice-replication.md).
- [ ] Updated model to `gemini-3.8-flash-tts` (creative fidelity) or `gemini-3.8-flash-lite-tts` (fast/cost-efficient replacement for `gemini-3.1-flash-tts-preview`)
- [ ] Moved turn-level directions (`style`) and speaker labels (`speaker`) out of transcript `text` into `annotations: [{"type": "speech_metadata", "speaker": "...", "style": "..."}]` (`text` is treated strictly as a verbatim transcript)
- [ ] Specified `speaker` inside `speech_metadata` on **every** turn in multi-speaker requests
- [ ] Accounted for default WAV (`audio/wav` with RIFF header) output on unary requests: remove manual WAV header wrappers (`wave` / `ffmpeg`), or explicitly set `response_format={"type": "audio", "mime_type": "audio/l16"}` (or `"audio/mulaw"` / `"audio/alaw"`). Streaming (`stream=True`) still defaults to headerless `audio/l16`.
- [ ] Kept inline tags (`<laugh>`, `<sigh>`, `<cough>`, `<breath>`, `<short pause>`) and pipe backchannels (`|mhm|`) only for point-in-time vocal events; avoid sound-effect tags
- [ ] Replaced multi-paragraph `"Audio Profile"` / `"Director's Notes"` blocks with a custom voice created via [Voice Design](https://aistudio.google.com/docs/voice-design.md) (`client.voices.create()` → `voice_...`)
### Migrate to Gemini Nano Banana 2.1 (`gemini-nano-banana-2.1`)
For full image generation details, resolutions, aspect ratios, and multi-image referencing, fetch the [Image Generation guide](https://aistudio.google.com/docs/image-generation.md) and [Gemini Nano Banana 2.1 model page](https://aistudio.google.com/docs/models/gemini-nano-banana-2.1.md).
- [ ] Updated model name to `gemini-nano-banana-2.1`
- [ ] Take advantage of expanded 131,072 input token limit (65k on previous generation) and 32,768 output token limit
- [ ] Multi-image reference support: up to 14 reference images (character consistency for up to 4 characters, object fidelity for up to 10 objects)
- [ ] Support for wide / panoramic aspect ratios (`1:4`, `4:1`, `1:8`, `8:1`) and output resolutions (`1K` default, `2K`, `4K`)
- [ ] Grounding with Google Web and Image Search supported natively
- [ ] Configurable thinking levels (`minimal`, `medium` default, `high`)
### Migrate to Gemini Omni 1.2 Flash (`gemini-omni-1.2-flash`)
For full video generation, editing, extension, and reference parameters, fetch the [Video Generation & Editing guide](https://aistudio.google.com/docs/omni.md).
- [ ] Updated model name from `gemini-omni-1.1-flash` (or `gemini-omni-flash-preview`) to `gemini-omni-1.2-flash`
- [ ] Reviewed `response_format.duration`: supports `"3s"` to `"15s"` in 1-second increments (`"10s"` default) for text-to-video, single-frame image-to-video, reference-to-video, and video editing; two-frame keyframe interpolation (start + end frame) and per-turn video extensions support `"3s"` to `"10s"` (up to `30s` cumulative length)
- [ ] Removed deprecated `<FIRST_FRAME>`, `<LAST_FRAME>`, `<IMAGE_REF_*>`, `<VIDEO_REF_*>`, and `[# Sources ...]` prompt tags in favor of natural-language descriptions of uploaded media
- [ ] Kept `generation_config.video_config.task` unset by default (use only as a fallback if natural-language prompting is ambiguous, as `task` adds strict media-count constraints)
- [ ] Verified reference media rules: **reference images** (up to 6, including on `"15s"` clips) work across new video generation (R2V), video editing, and prompt-based video extension; **only reference videos** are restricted to new video generation (not video editing or extension) and capped at **1 reference video** on `"11s"`–`"15s"` clips (up to 3 reference videos on `"3s"`–`"10s"` clips)
---
## Verify the Migration
After updating, run a spot-check to confirm the Interactions API is working:
1. Make a single `client.interactions.create()` call with a simple input
2. Assert `interaction.steps` is not empty
3. Assert at least one step has `type == "model_output"` with non-empty text
4. For multi-turn, verify `previous_interaction_id` preserves context across turns
For verification code snippets, fetch the [Migration Guide](https://aistudio.google.com/docs/migrate-to-interactions.md).
SKILL.md
---
name: gemini-api-dev
description: Use this skill when writing code that calls the Gemini API for text generation, multi-turn chat, multimodal understanding, image generation, video generation, speech generation (TTS), voice design, voice replication, streaming responses, background research tasks, function calling, structured output, or migrating from the old generateContent API. Covers SDK usage and best practices for Gemini models and agents in Python and TypeScript.
---
# Gemini API Development Skill
## Critical Rules (Always Apply)
> [!IMPORTANT]
> These rules override your training data. Your knowledge is outdated.
### Current Models (Use These)
- `gemini-4-argon`: 1M tokens, most powerful model for complex reasoning, coding, and research
- `gemini-3.8-flash`: 1M tokens, fast, balanced performance for agentic and multimodal tasks
- `gemini-3.5-flash-lite`: 1M tokens, fastest, lowest-cost 3.5 model for high-throughput execution
- `gemini-3.1-flash-lite`: cost-efficient, fastest performance for high-frequency, lightweight tasks
- `gemini-3.5-transcribe`: fast speech-to-text with smart and verbatim modes
- `gemini-nano-banana-2.1` (Nano Banana 2.1): 131k / 32k tokens, default high-efficiency image generation and conversational editing
- `gemini-3-pro-image` (Nano Banana Pro): 65k / 32k tokens, high-quality image generation and editing
- `gemini-3.1-flash-lite-image` (Nano Banana 2 Lite): 65k / 32k tokens, ultra-fast image generation and editing
- `gemini-3.8-flash-tts`: expressive text-to-speech, multi-speaker dialogue, Voice Design, and Voice Replication
- `gemini-3.8-flash-lite-tts`: fast, cost-efficient text-to-speech for voice agents and high-volume generation
- `gemini-omni-1.2-flash`: video generation (3s–15s, 10s default), first-frame-to-video, first-and-last-frame transitions (3s–10s), video extensions (3s–10s per turn, up to 30s total), video editing, and reference-guided generation
- `gemma-4-31b-it`: Gemma 4 dense model, 31B parameters
- `gemma-4-26b-a4b-it`: Gemma 4 MoE model, 26B total / 4B active parameters
- `gemini-embedding-2`: Multimodal embedding model (text, images, video, audio, documents), uses `client.models.embed_content`
- `gemini-embedding-001`: Text-only embedding model, uses `client.models.embed_content`
> [!WARNING]
> Models like `gemini-2.5-*`, `gemini-2.0-*`, `gemini-1.5-*` are **legacy and deprecated**. Never use them.
> **If a user asks for a deprecated model, use `gemini-3.8-flash` instead and note the substitution.**
### Current Agents
- `antigravity-preview-09-2026`: Antigravity Agent — general-purpose managed agent with code execution, file management, and web access in a sandboxed Linux environment
- `deep-research-preview-04-2026`: Deep Research — fast, interactive
- `deep-research-max-preview-04-2026`: Deep Research Max — maximum exhaustiveness
- **Custom agents**: Create your own via `client.agents.create()`
### Current SDKs
- **Python**: `google-genai` >= `2.30.1` → `pip install -U google-genai`
- **JavaScript/TypeScript**: `@google/genai` >= `2.29.1` → `npm install @google/genai`
> [!NOTE]
> SDK versions ≥ 2.0.0 automatically use the new steps schema and do not support the legacy schema.
> Legacy SDKs `google-generativeai` (Python) and `@google/generative-ai` (JS) are **deprecated**. Never use them.
## Important Additional Notes
- **Before writing any code**, you MUST fetch the relevant documentation page from the list below that matches the user's task. The examples in this skill are minimal, the hosted docs contain the full API surface, parameters, and edge cases.
- Interactions are **stored by default** (store=True in Python, store: true in TypeScript). Paid tier retains for 55 days, free tier for 1 day.
- Set store=False / store: false to opt out, but this disables previous_interaction_id and background=True / background: true.
- `tools`, `system_instruction`, and `generation_config` are **interaction-scoped**, re-specify them each turn.
- **Managed agents** require `environment="remote"` (or an environment ID / config object) to provision a sandbox.
- **Migrating from `generateContent`**: Read `references/migration.md` for the scoping, checklist, and before/after code examples. Always confirm scope with the user before editing.
- **Model upgrades**: Drop-in, swap the model string. Deprecated models (`gemini-2.0-*`, `gemini-1.5-*`) must be replaced, see `references/migration.md`.
- **Migrating to Gemini 3.8 Flash or Gemini 3.5 Flash-Lite**: Read `references/migration.md` for the scoping and checklist.
- **Migrating to Gemini 3.8 TTS (`gemini-3.8-flash-tts` / `gemini-3.8-flash-lite-tts`)**: Read `references/migration.md` for breaking changes from `gemini-3.1-flash-tts-preview` (`speech_metadata` annotations, inline vocal tags, default WAV `audio/wav` unary output vs `audio/l16` streaming output, and Voice Design personas).
- **Migrating to Gemini Nano Banana 2.1 (`gemini-nano-banana-2.1`)**: Read `references/migration.md` for upgrading from `gemini-3.1-flash-image` (deprecated) and using multi-image reference fusion with up to 14 reference images.
- **Migrating to Gemini Omni 1.2 Flash (`gemini-omni-1.2-flash`)**: Read `references/migration.md` for upgrading from `gemini-omni-1.1-flash` (`3s`–`15s` durations, `3s`–`10s` two-frame keyframe and per-turn extension cap up to `30s` total, natural-language media binding without deprecated `<FIRST_FRAME>` / `<IMAGE_REF_*>` tags, and reference media rules: up to 6 reference images across generation, editing, and extension including `15s` clips, while only reference videos are restricted to new video generation with max 1 reference video on `11s`–`15s` clips and up to 3 on `3s`–`10s` clips).
## Quick Start
### Python
```python
from google import genai
client = genai.Client()
interaction = client.interactions.create(
model="gemini-3.8-flash",
input="Tell me a short joke about programming."
)
print(interaction.output_text)
```
### JavaScript/TypeScript
```typescript
import { GoogleGenAI } from "@google/genai";
const client = new GoogleGenAI({});
const interaction = await client.interactions.create({
model: "gemini-3.8-flash",
input: "Tell me a short joke about programming.",
});
console.log(interaction.output_text);
```
## Response Helpers
The SDK provides convenience properties on the `Interaction` response object to simplify common access patterns:
| Property | Type | Description |
|---|---|---|
| `output_text` | `string \| null` | The last consecutive run of text from the trailing `model_output` steps. Returns the combined text when the model's final output contains multiple text parts. |
| `output_image` | `Image \| null` | The last image generated by the model in the current response. Returns an object with `data` (base64) and `mime_type`. |
| `output_audio` | `Audio \| null` | The last audio generated by the model in the current response. Returns an object with `data` (base64) and `mime_type`. |
## Stateful Conversation
### Python
```python
interaction1 = client.interactions.create(
model="gemini-3.8-flash",
input="Hi, my name is Phil."
)
# Second turn — server remembers context
interaction2 = client.interactions.create(
model="gemini-3.8-flash",
input="What is my name?",
previous_interaction_id=interaction1.id
)
print(interaction2.output_text)
```
### JavaScript/TypeScript
```typescript
const interaction1 = await client.interactions.create({
model: "gemini-3.8-flash",
input: "Hi, my name is Phil.",
});
const interaction2 = await client.interactions.create({
model: "gemini-3.8-flash",
input: "What is my name?",
previous_interaction_id: interaction1.id,
});
console.log(interaction2.output_text);
```
## Deep Research Agent
Use `deep-research-preview-04-2026` for fast research or `deep-research-max-preview-04-2026` for maximum exhaustiveness. Agents require `background=True`.
### Python
```python
import time
interaction = client.interactions.create(
agent="deep-research-preview-04-2026",
input="Research the history of Google TPUs.",
background=True
)
while True:
interaction = client.interactions.get(interaction.id)
if interaction.status == "completed":
print(interaction.output_text)
break
elif interaction.status == "failed":
print(f"Failed: {interaction.error}")
break
time.sleep(10)
```
### JavaScript/TypeScript
```typescript
import { GoogleGenAI } from "@google/genai";
const client = new GoogleGenAI({});
// Start background research
const initialInteraction = await client.interactions.create({
agent: "deep-research-preview-04-2026",
input: "Research the history of Google TPUs.",
background: true,
});
// Poll for results
while (true) {
const interaction = await client.interactions.get(initialInteraction.id);
if (interaction.status === "completed") {
console.log(interaction.output_text);
break;
} else if (["failed", "cancelled"].includes(interaction.status)) {
console.log(`Failed: ${interaction.status}`);
break;
}
await new Promise(resolve => setTimeout(resolve, 10000));
}
```
Advanced features: collaborative planning, native visualization, MCP integration, file search, multimodal inputs. See [Deep Research docs](https://aistudio.google.com/docs/deep-research.md).
## Managed Agents
Managed agents run inside a sandboxed Linux environment hosted by Google. Fetch the [Managed Agents Quickstart](https://aistudio.google.com/docs/managed-agents-quickstart.md) before writing agent code.
### Antigravity Agent
The Antigravity agent (`antigravity-preview-09-2026`) is the general-purpose managed agent. It can execute code (Bash, Python, Node.js), manage files, browse the web, and use Google Search. It uses `gemini-3.8-flash` by default and also supports `gemini-4-argon` via `agent_config={"type": "antigravity", "model": "gemini-4-argon"}`. See [Antigravity Agent docs](https://aistudio.google.com/docs/antigravity-agent.md) for capabilities, tools, multimodal input, and pricing.
#### Python
```python
from google import genai
client = genai.Client()
interaction = client.interactions.create(
agent="antigravity-preview-09-2026",
input="Write a Python script that generates the first 20 Fibonacci numbers and saves them to fibonacci.txt. Then read the file and print its contents.",
environment="remote",
)
print(f"Environment ID: {interaction.environment_id}")
print(interaction.output_text)
```
#### JavaScript/TypeScript
```typescript
import { GoogleGenAI } from "@google/genai";
const client = new GoogleGenAI({});
const interaction = await client.interactions.create({
agent: "antigravity-preview-09-2026",
input: "Write a Python script that generates the first 20 Fibonacci numbers and saves them to fibonacci.txt. Then read the file and print its contents.",
environment: "remote",
});
console.log(`Environment ID: ${interaction.environment_id}`);
console.log(interaction.output_text);
```
### Custom Agents
See [Building Custom Agents docs](https://aistudio.google.com/docs/custom-agents.md).
#### Python
```python
agent = client.agents.create(
id="code-reviewer",
base_agent="antigravity-preview-09-2026",
system_instruction="You are a senior code reviewer. Check every file for bugs, style issues, and security vulnerabilities.",
base_environment={
"type": "remote",
"sources": [
{
"type": "repository",
"source": "https://github.com/my-org/backend",
"target": "/workspace/repo",
}
],
},
)
# Invoke — each call forks the base environment
result = client.interactions.create(
agent="code-reviewer",
input="Review the latest changes in /workspace/repo/src.",
environment="remote",
)
print(result.output_text)
```
#### JavaScript/TypeScript
```typescript
const agent = await client.agents.create({
id: "code-reviewer",
base_agent: "antigravity-preview-09-2026",
system_instruction: "You are a senior code reviewer. Check every file for bugs, style issues, and security vulnerabilities.",
base_environment: {
type: "remote",
sources: [
{
type: "repository",
source: "https://github.com/my-org/backend",
target: "/workspace/repo",
}
],
},
});
const result = await client.interactions.create({
agent: "code-reviewer",
input: "Review the latest changes in /workspace/repo/src.",
environment: "remote",
});
console.log(result.output_text);
```
Manage agents with `client.agents.list()`, `client.agents.get(id=...)`, and `client.agents.delete(id=...)`.
## Streaming
Set `stream=True` to receive incremental server-sent events. Each stream follows: `interaction.created` → (`step.start` → `step.delta`(s) → `step.stop`)+ → `interaction.completed`.
### Python
```python
for event in client.interactions.create(
model="gemini-3.8-flash",
input="Explain quantum entanglement in simple terms.",
stream=True,
):
if event.event_type == "step.delta":
if event.delta.type == "text":
print(event.delta.text, end="", flush=True)
elif event.event_type == "interaction.completed":
print(f"\n\nTotal Tokens: {event.interaction.usage.total_tokens}")
```
### JavaScript/TypeScript
```typescript
const stream = await client.interactions.create({
model: "gemini-3.8-flash",
input: "Explain quantum entanglement in simple terms.",
stream: true,
});
for await (const event of stream) {
if (event.event_type === "step.delta") {
if (event.delta.type === "text") {
process.stdout.write(event.delta.text);
}
} else if (event.event_type === "interaction.completed") {
console.log(`\n\nTotal Tokens: ${event.interaction?.usage?.total_tokens}`);
}
}
```
For streaming with tools, thinking, agents, and image generation see the full [Streaming guide](https://aistudio.google.com/docs/streaming.md).
## Documentation Pages
**You MUST fetch the matching page below before writing code.** These hosted docs are the source of truth for parameters, types, and edge cases — do not rely solely on the examples above.
**Core Documentation:**
- [Interactions API Overview](https://aistudio.google.com/docs/interactions-overview.md)
- [Quickstart](https://aistudio.google.com/docs/get-started.md)
- [Text Generation](https://aistudio.google.com/docs/text-generation.md)
- [Streaming](https://aistudio.google.com/docs/streaming.md)
- [Tokens](https://aistudio.google.com/docs/tokens.md)
- [API Keys](https://aistudio.google.com/docs/api-key.md)
**Tools & Function Calling:**
- [Function Calling](https://aistudio.google.com/docs/function-calling.md)
- [Google Search](https://aistudio.google.com/docs/google-search.md)
- [Code Execution](https://aistudio.google.com/docs/code-execution.md)
- [URL Context](https://aistudio.google.com/docs/url-context.md)
- [File Search](https://aistudio.google.com/docs/file-search.md)
- [Tool Combination](https://aistudio.google.com/docs/tool-combination.md)
- [Computer Use](https://aistudio.google.com/docs/computer-use.md)
- [Maps Grounding](https://aistudio.google.com/docs/maps-grounding.md)
**Generation & Output:**
- [Structured Output](https://aistudio.google.com/docs/structured-output.md)
- [Thinking](https://aistudio.google.com/docs/thinking.md)
- [Thought Signatures](https://aistudio.google.com/docs/thought-signatures.md)
- [Image Generation](https://aistudio.google.com/docs/image-generation.md)
- [Image Understanding](https://aistudio.google.com/docs/image-understanding.md)
- [Video Generation & Editing (Omni Flash)](https://aistudio.google.com/docs/omni.md)
- [Speech Generation (TTS)](https://aistudio.google.com/docs/speech-generation.md)
- [Voice Design](https://aistudio.google.com/docs/voice-design.md)
- [Voice Replication](https://aistudio.google.com/docs/voice-replication.md)
- [Music Generation](https://aistudio.google.com/docs/music-generation.md)
- [Embeddings](https://aistudio.google.com/docs/embeddings.md)
**Multimodal Understanding:**
- [Audio](https://aistudio.google.com/docs/audio.md)
- [Audio Transcription](https://aistudio.google.com/docs/transcribe.md)
- [Video Understanding](https://aistudio.google.com/docs/video-understanding.md)
- [Document Processing](https://aistudio.google.com/docs/document-processing.md)
**Files & Context:**
- [Files](https://aistudio.google.com/docs/files.md)
- [File Input Methods](https://aistudio.google.com/docs/file-input-methods.md)
- [Long Context](https://aistudio.google.com/docs/long-context.md)
- [Caching](https://aistudio.google.com/docs/caching.md)
- [Media Resolution](https://aistudio.google.com/docs/media-resolution.md)
**Agents:**
- [Agents Overview](https://aistudio.google.com/docs/agents.md)
- [Managed Agents Quickstart](https://aistudio.google.com/docs/managed-agents-quickstart.md)
- [Antigravity Agent](https://aistudio.google.com/docs/antigravity-agent.md)
- [Agent Environments](https://aistudio.google.com/docs/agent-environment.md)
- [Agent Hooks](https://aistudio.google.com/docs/agent-hooks.md)
- [Agent Credentials](https://aistudio.google.com/docs/agent-credentials.md)
- [Building Custom Agents](https://aistudio.google.com/docs/custom-agents.md)
- [Deep Research](https://aistudio.google.com/docs/deep-research.md)
**Advanced Features:**
- [Latest Model (Gemini 4 Argon)](https://aistudio.google.com/docs/latest-model.md)
- [Background Execution](https://aistudio.google.com/docs/background-execution.md)
- [Flex Inference](https://aistudio.google.com/docs/flex-inference.md)
- [Priority Inference](https://aistudio.google.com/docs/priority-inference.md)
**API Reference:**
- [API Reference](https://aistudio.google.com/api-reference/interactions-api.md)
- [OpenAPI Spec](https://aistudio.google.com/api-reference/interactions.openapi.json)
- [May 2026 Breaking Changes Migration Guide](https://aistudio.google.com/docs/interactions-breaking-changes-may-2026.md)
## Data Model
An `Interaction` response contains `steps`, an array of typed step objects representing a structured timeline of the interaction turn.
### Step Types
**User steps:**
- `user_input`: User input (text, audio, multimodal). Contains `content` array.
**Model/server steps:**
- `model_output`: Final model generation. Contains `content` array with `text`, `image`, `audio`, etc.
- `thought`: Model reasoning/Chain of Thought. Has `signature` field (required) and optional `summary`.
- `function_call`: Tool call request (`id`, `name`, `arguments`).
- `function_result`: Tool result you send back (`call_id`, `name`, `result`).
- `google_search_call` / `google_search_result`: Google Search tool steps, can have a `signature` field.
- `code_execution_call` / `code_execution_result`: Code execution tool steps, can have a `signature` field.
- `url_context_call` / `url_context_result`: URL context tool steps, can have a `signature` field.
- `mcp_server_tool_call` / `mcp_server_tool_result`: Remote MCP tool steps.
- `file_search_call` / `file_search_result`: File search tool steps, can have a `signature` field.
### Content types (inside `content` array on `model_output` and `user_input` steps)
- `text`: Text content (`text` field, plus optional `annotations` such as `{"type": "speech_metadata", "speaker": "...", "style": "..."}` for TTS)
- `image` / `audio` / `document` / `video`: Content with `data`, `mime_type`, or `uri`
### Streaming Event Types
| Event | Description |
|---|---|
| `interaction.created` | Interaction created; includes metadata. |
| `interaction.status_update` | Interaction-level status change. |
| `step.start` | A new step begins. Contains step `type` and initial metadata. |
| `step.delta` | Incremental data for the current step. Contains a typed `delta` object. |
| `step.stop` | The step is complete. Contains `index`. |
| `interaction.completed` | Interaction finished. Contains final `usage`. |
### Delta Types
| Delta Type | Parent Step | Description |
|---|---|---|
| `text` | `model_output` | Incremental text token. |
| `audio` | `model_output` | audio chunk (base64). |
| `image` | `model_output` | image chunk (base64). |
| `thought_summary` | `thought` | thinking summary text. |
| `thought_signature` | `thought` | Opaque signature for thought verification. |
**Status values:** `completed`, `in_progress`, `queued`, `requires_action`, `incomplete`, `failed`, `cancelled`
## Gemini Live API
For real-time, bidirectional audio/video/text streaming with the Gemini Live API (`gemini-3.8-live`, `gemini-3.8-live-extended-thinking`, and `gemini-3.5-transcribe-live`), install the **`google-gemini/gemini-live-api-dev`** skill. It covers WebSocket streaming, voice activity detection, background reasoning (extended thinking), asynchronous function calling, session management, ephemeral tokens, and more.