Skip to content

Automate browser based workflows with AI

feat: modernize LLM models and clean up docs (#5684)

Suchintan539ffd7e4eb4
diff --git a/.env.example b/.env.example
index 3405d14b4..071549d9e 100644
--- a/.env.example
+++ b/.env.example
@@ -60,23 +60,11 @@ ENABLE_GEMINI=false
# GEMINI_API_KEY: Your Gemini API key for accessing models like Gemini 2.5 Pro.
GEMINI_API_KEY=""
-# ENABLE_NOVITA: Set to true to enable Novita AI as a language model provider.
-ENABLE_NOVITA=false
-# NOVITA_API_KEY: Your Novita AI API key.
-NOVITA_API_KEY=""
-
-# ENABLE_VOLCENGINE: Set to true to enable Volcengine(ByteDance Doubao) as a language model provider.
-ENABLE_VOLCENGINE=false
-# VOLCENGINE_API_KEY: Your Volcengine(ByteDance Doubao) API key.
-VOLCENGINE_API_KEY=""
-# VOLCENGINE_API_BASE: The base URL for Volcengine(ByteDance Doubao) API.
-VOLCENGINE_API_BASE="https://ark.cn-beijing.volces.com/api/v3"
-
# LLM_KEY: The chosen language model to use. This should be one of the models
-# provided by the enabled LLM providers (e.g., OPENAI_GPT5, OPENAI_GPT5_MINI, OPENAI_GPT5_NANO,
-# AZURE_OPENAI_GPT5, AZURE_OPENAI_GPT5_MINI, AZURE_OPENAI_GPT5_NANO, ANTHROPIC_CLAUDE4.5_SONNET,
-# ANTHROPIC_CLAUDE4_SONNET, ANTHROPIC_CLAUDE3.7_SONNET, OPENAI_GPT4_TURBO,
-# OPENAI_GPT4V, ANTHROPIC_CLAUDE3, AZURE_OPENAI_GPT4V).
+# provided by the enabled LLM providers (e.g., OPENAI_GPT5_5, OPENAI_GPT5_4,
+# ANTHROPIC_CLAUDE4.7_OPUS, ANTHROPIC_CLAUDE4.6_SONNET, GEMINI_3_PRO,
+# BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE).
+# See docs: https://www.skyvern.com/docs/self-hosted/llm-configuration
LLM_KEY=""
# a cheaper LLM providers to help finishing some small tasks, like custom selection or svg conversion. If empty, it will be the same as LLM_KEY
SECONDARY_LLM_KEY=""
diff --git a/README.md b/README.md
index 68dc9a454..72fe2505e 100644
--- a/README.md
+++ b/README.md
@@ -232,32 +232,40 @@ print(task)
## Advanced Usage
### Control your own browser (Chrome)
-> [!WARNING]
-> Since [Chrome 136](https://developer.chrome.com/blog/remote-debugging-port), Chrome refuses any CDP connect to the browser using the default user_data_dir. In order to use your browser data, Skyvern copies your default user_data_dir to `./tmp/user_data_dir` the first time connecting to your local browser.
-1. Just With Python Code
+Let Skyvern control your existing Chrome browser — with all your cookies, logins, and extensions.
+
+#### Step 1: Enable remote debugging in Chrome
+
+1. Open Chrome and navigate to `chrome://inspect/#remote-debugging`
+2. Click **Enable** to start the debugging server
+3. You should see: **Server running at: 127.0.0.1:9222**
+
+> [!TIP]
+> The `skyvern init browser` command can do this automatically — it opens `chrome://inspect/#remote-debugging`, waits for you to enable it, and saves the config.
+
+#### Step 2: Connect Skyvern
+
+**Option A — Python Code:**
```python
from skyvern import Skyvern
-# The path to your Chrome browser. This example path is for Mac.
-browser_path = "/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
skyvern = Skyvern(
base_url="http://localhost:8000",
api_key="YOUR_API_KEY",
- browser_path=browser_path,
+ browser_address="http://127.0.0.1:9222",
)
task = await skyvern.run_task(
prompt="Find the top post on hackernews today",
)
```
-2. With Skyvern Service
+**Option B — Skyvern Service:**
Add two variables to your .env file:
```bash
-# The path to your Chrome browser. This example path is for Mac.
-CHROME_EXECUTABLE_PATH="/Applications/Google Chrome.app/Contents/MacOS/Google Chrome"
BROWSER_TYPE=cdp-connect
+BROWSER_REMOTE_DEBUGGING_URL=http://127.0.0.1:9222
```
Restart Skyvern service `skyvern run all` and run the task through UI or code
@@ -436,7 +444,7 @@ Skyvern currently supports the following password manager integrations:
Skyvern supports the Model Context Protocol (MCP) to allow you to use any LLM that supports MCP.
-See the MCP documentation [here](https://github.com/Skyvern-AI/skyvern/blob/main/skyvern/cli/mcp_tools/README.md)
+See the MCP documentation [here](https://www.skyvern.com/docs/integrations/mcp#mcp-server)
## Zapier / Make.com / N8N Integration
Skyvern supports Zapier, Make.com, and N8N to allow you to connect your Skyvern workflows to other apps.
@@ -516,124 +524,16 @@ More extensive documentation can be found on our [📕 docs page](https://www.sk
# Supported LLMs
| Provider | Supported Models |
| -------- | ------- |
-| OpenAI | GPT-5, GPT-5.2, GPT-4.1, o3, o4-mini |
-| Anthropic | Claude 4 (Sonnet, Opus), Claude 4.5 (Haiku, Sonnet, Opus) |
-| Azure OpenAI | Any GPT models. Better performance with a multimodal llm (azure/gpt4-o) |
-| AWS Bedrock | Claude 3.5, Claude 3.7, Claude 4 (Sonnet, Opus), Claude 4.5 (Sonnet, Opus) |
-| Gemini | Gemini 3 Pro/Flash, Gemini 2.5 Pro/Flash |
+| OpenAI | GPT-5.5, GPT-5.4, GPT-5, GPT-4.1, o3, o4-mini |
+| Anthropic | Claude 4.7 Opus, Claude 4.6 (Sonnet, Opus), Claude 4.5 (Haiku, Sonnet, Opus) |
+| Azure OpenAI | Any GPT models deployed to your Azure subscription |
+| AWS Bedrock | Claude 4.7, Claude 4.6 (Sonnet, Opus), Claude 4.5 (Sonnet, Opus) |
+| Gemini | Gemini 3.1 Pro, Gemini 3 Flash, Gemini 2.5 Pro/Flash |
| Ollama | Run any locally hosted model via [Ollama](https://github.com/ollama/ollama) |
| OpenRouter | Access models through [OpenRouter](https://openrouter.ai) |
| OpenAI-compatible | Any custom API endpoint that follows OpenAI's API format (via [liteLLM](https://docs.litellm.ai/docs/providers/openai_compatible)) |
-#### Environment Variables
-
-##### OpenAI
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_OPENAI`| Register OpenAI models | Boolean | `true`, `false` |
-| `OPENAI_API_KEY` | OpenAI API Key | String | `sk-1234567890` |
-| `OPENAI_API_BASE` | OpenAI API Base, optional | String | `https://openai.api.base` |
-| `OPENAI_ORGANIZATION` | OpenAI Organization ID, optional | String | `your-org-id` |
-
-Recommended `LLM_KEY`: `OPENAI_GPT5`, `OPENAI_GPT5_2`, `OPENAI_GPT4_1`, `OPENAI_O3`, `OPENAI_O4_MINI`
-
-##### Anthropic
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_ANTHROPIC` | Register Anthropic models| Boolean | `true`, `false` |
-| `ANTHROPIC_API_KEY` | Anthropic API key| String | `sk-1234567890` |
-
-Recommended `LLM_KEY`: `ANTHROPIC_CLAUDE4.5_OPUS`, `ANTHROPIC_CLAUDE4.5_SONNET`, `ANTHROPIC_CLAUDE4_OPUS`, `ANTHROPIC_CLAUDE4_SONNET`
-
-##### Azure OpenAI
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_AZURE` | Register Azure OpenAI models | Boolean | `true`, `false` |
-| `AZURE_API_KEY` | Azure deployment API key | String | `sk-1234567890` |
-| `AZURE_DEPLOYMENT` | Azure OpenAI Deployment Name | String | `skyvern-deployment`|
-| `AZURE_API_BASE` | Azure deployment api base url| String | `https://skyvern-deployment.openai.azure.com/`|
-| `AZURE_API_VERSION` | Azure API Version| String | `2024-02-01`|
-
-Recommended `LLM_KEY`: `AZURE_OPENAI`
-
-##### AWS Bedrock
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_BEDROCK` | Register AWS Bedrock models. To use AWS Bedrock, you need to make sure your [AWS configurations](https://github.com/boto/boto3?tab=readme-ov-file#using-boto3) are set up correctly first. | Boolean | `true`, `false` |
-
-Recommended `LLM_KEY`: `BEDROCK_ANTHROPIC_CLAUDE4.5_OPUS_INFERENCE_PROFILE`, `BEDROCK_ANTHROPIC_CLAUDE4.5_SONNET_INFERENCE_PROFILE`, `BEDROCK_ANTHROPIC_CLAUDE4_OPUS_INFERENCE_PROFILE`
-
-##### Gemini
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_GEMINI` | Register Gemini models| Boolean | `true`, `false` |
-| `GEMINI_API_KEY` | Gemini API Key| String | `your_google_gemini_api_key`|
-
-Recommended `LLM_KEY`: `GEMINI_3.0_FLASH`, `GEMINI_2.5_PRO`, `GEMINI_2.5_FLASH`, `GEMINI_2.5_PRO_PREVIEW`, `GEMINI_2.5_FLASH_PREVIEW`
-
-##### Ollama
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_OLLAMA`| Register local models via Ollama | Boolean | `true`, `false` |
-| `OLLAMA_SERVER_URL` | URL for your Ollama server | String | `http://host.docker.internal:11434` |
-| `OLLAMA_MODEL` | Ollama model name to load | String | `qwen2.5:7b-instruct` |
-| `OLLAMA_SUPPORTS_VISION` | Enable vision support | Boolean | `true`, `false` |
-
-Recommended `LLM_KEY`: `OLLAMA`
-
-Note: Set `OLLAMA_SUPPORTS_VISION=true` for vision models like qwen3-vl, llava, etc.
-
-##### OpenRouter
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_OPENROUTER`| Register OpenRouter models | Boolean | `true`, `false` |
-| `OPENROUTER_API_KEY` | OpenRouter API key | String | `sk-1234567890` |
-| `OPENROUTER_MODEL` | OpenRouter model name | String | `mistralai/mistral-small-3.1-24b-instruct` |
-| `OPENROUTER_API_BASE` | OpenRouter API base URL | String | `https://api.openrouter.ai/v1` |
-
-Recommended `LLM_KEY`: `OPENROUTER`
-
-##### OpenAI-Compatible
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `ENABLE_OPENAI_COMPATIBLE`| Register a custom OpenAI-compatible API endpoint | Boolean | `true`, `false` |
-| `OPENAI_COMPATIBLE_MODEL_NAME` | Model name for OpenAI-compatible endpoint | String | `yi-34b`, `gpt-3.5-turbo`, `mistral-large`, etc.|
-| `OPENAI_COMPATIBLE_API_KEY` | API key for OpenAI-compatible endpoint | String | `sk-1234567890`|
-| `OPENAI_COMPATIBLE_API_BASE` | Base URL for OpenAI-compatible endpoint | String | `https://api.together.xyz/v1`, `http://localhost:8000/v1`, etc.|
-| `OPENAI_COMPATIBLE_API_VERSION` | API version for OpenAI-compatible endpoint, optional| String | `2023-05-15`|
-| `OPENAI_COMPATIBLE_MAX_TOKENS` | Maximum tokens for completion, optional| Integer | `4096`, `8192`, etc.|
-| `OPENAI_COMPATIBLE_TEMPERATURE` | Temperature setting, optional| Float | `0.0`, `0.5`, `0.7`, etc.|
-| `OPENAI_COMPATIBLE_SUPPORTS_VISION` | Whether model supports vision, optional| Boolean | `true`, `false`|
-
-Supported LLM Key: `OPENAI_COMPATIBLE`
-
-##### General LLM Configuration
-| Variable | Description| Type | Sample Value|
-| -------- | ------- | ------- | ------- |
-| `LLM_KEY` | The name of the model you want to use | String | See supported LLM keys above |
-| `SECONDARY_LLM_KEY` | The name of the model for mini agents skyvern runs with | String | See supported LLM keys above |
-| `LLM_CONFIG_MAX_TOKENS` | Override the max tokens used by the LLM | Integer | `128000` |
-
-# Feature Roadmap
-This is our planned roadmap for the next few months. If you have any suggestions or would like to see a feature added, please don't hesitate to reach out to us [via email](mailto:[email protected]) or [discord](https://discord.gg/fG2XXEuQX3).
-
-- [x] **Open Source** - Open Source Skyvern's core codebase
-- [x] **Workflow support** - Allow support to chain multiple Skyvern calls together
-- [x] **Improved context** - Improve Skyvern's ability to understand content around interactable elements by introducing feeding relevant label context through the text prompt
-- [x] **Cost Savings** - Improve Skyvern's stability and reduce the cost of running Skyvern by optimizing the context tree passed into Skyvern
-- [x] **Self-serve UI** - Deprecate the Streamlit UI in favour of a React-based UI component that allows users to kick off new jobs in Skyvern
-- [x] **Workflow UI Builder** - Introduce a UI to allow users to build and analyze workflows visually
-- [x] **Chrome Viewport streaming** - Introduce a way to live-stream the Chrome viewport to the user's browser (as a part of the self-serve UI)
-- [x] **Past Runs UI** - Deprecate the Streamlit UI in favour of a React-based UI that allows you to visualize past runs and their results
-- [X] **Auto workflow builder ("Observer") mode** - Allow Skyvern to auto-generate workflows as it's navigating the web to make it easier to build new workflows
-- [x] **Prompt Caching** - Introduce a caching layer to the LLM calls to dramatically reduce the cost of running Skyvern (memorize past actions and repeat them!)
-- [x] **Web Evaluation Dataset** - Integrate Skyvern with public benchmark tests to track the quality of our models over time
-- [ ] **Improved Debug mode** - Allow Skyvern to plan its actions and get "approval" before running them, allowing you to debug what it's doing and more easily iterate on the prompt
-- [ ] **Chrome Extension** - Allow users to interact with Skyvern through a Chrome extension (incl voice mode, saving tasks, etc.)
-- [ ] **Skyvern Action Recorder** - Allow Skyvern to watch a user complete a task and then automatically generate a workflow for it
-- [ ] **Interactable Livestream** - Allow users to interact with the livestream in real-time to intervene when necessary (such as manually submitting sensitive forms)
-- [ ] **Integrate LLM Observability tools** - Integrate LLM Observability tools to allow back-testing prompt changes with specific data sets + visualize the performance of Skyvern over time
-- [x] **Langchain Integration** - Create langchain integration in langchain_community to use Skyvern as a "tool".
+For detailed LLM configuration including all available model keys, environment variables, and multi-model setups, see the [LLM Configuration docs](https://www.skyvern.com/docs/self-hosted/llm-configuration).
# Contributing
diff --git a/docker-compose.yml b/docker-compose.yml
index eda6ee665..3b835a3cd 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -40,78 +40,65 @@ services:
- DATABASE_STRING=postgresql+psycopg://skyvern:skyvern@postgres:5432/skyvern
- BROWSER_TYPE=chromium-headful
- ENABLE_CODE_BLOCK=true
+ # --- Control your own browser (Chrome) ---
+ # 1. Open Chrome and navigate to: chrome://inspect/#remote-debugging
+ # 2. Click "Enable" to start the debugging server
+ # 3. Uncomment the two lines below:
# - BROWSER_TYPE=cdp-connect
- # Use this command to start Chrome with remote debugging:
- # "C:\Program Files\Google\Chrome\Application\chrome.exe" --remote-debugging-port=9222 --user-data-dir="C:\chrome-cdp-profile" --no-first-run --no-default-browser-check
- # /Applications/Google\ Chrome.app/Contents/MacOS/Google\ Chrome --remote-debugging-port=9222 --user-data-dir="/Users/yourusername/chrome-cdp-profile" --no-first-run --no-default-browser-check
# - BROWSER_REMOTE_DEBUGGING_URL=http://host.docker.internal:9222/
# =========================
- # LLM Settings - Recommended to use skyvern CLI, `skyvern init llm` to setup your LLM's
+ # LLM Settings - Use `skyvern init llm` for interactive setup
# =========================
- # OpenAI Support:
- # If you want to use OpenAI as your LLM provider, uncomment the following lines and fill in your OpenAI API key.
+ # Docs: https://www.skyvern.com/docs/self-hosted/llm-configuration
+ #
+ # OpenAI:
# - ENABLE_OPENAI=true
- # - LLM_KEY=OPENAI_GPT4O
+ # - LLM_KEY=OPENAI_GPT5_5
# - OPENAI_API_KEY=<your_openai_key>
- # Gemini Support:
- # Gemini is a new LLM provider that is currently in beta. You can use it by uncommenting the following lines and filling in your Gemini API key.
- # - LLM_KEY=GEMINI
- # - ENABLE_GEMINI=true
- # - GEMINI_API_KEY=YOUR_GEMINI_KEY
- # - LLM_KEY=GEMINI_3.0_FLASH
- # If you want to use other LLM provider, like azure and anthropic:
+ #
+ # Anthropic:
# - ENABLE_ANTHROPIC=true
- # - LLM_KEY=ANTHROPIC_CLAUDE3.5_SONNET
+ # - LLM_KEY=ANTHROPIC_CLAUDE4.7_OPUS
# - ANTHROPIC_API_KEY=<your_anthropic_key>
- # Microsoft Azure OpenAI support:
- # If you'd like to use Microsoft Azure OpenAI as your managed LLM service integration with Skyvern, use the environment variables below.
- # In your Microsoft Azure subscription, you will need to provision the OpenAI service and deploy a model, in order to utilize it.
- # 1. Login to the Azure Portal
- # 2. Create an Azure Resource Group
- # 3. Create an OpenAI resource in the Resource Group (choose a region and pricing tier)
- # 4. From the OpenAI resource's Overview page, open the "Azure AI Foundry" portal (click the "Explore Azure AI Foundry Portal" button)
- # 5. In Azure AI Foundry, click "Shared Resources" --> "Deployments"
- # 6. Click "Deploy Model" --> "Deploy Base Model" --> select a model (specify this model "Deployment Name" value for the AZURE_DEPLOYMENT variable below)
+ #
+ # Gemini:
+ # - ENABLE_GEMINI=true
+ # - LLM_KEY=GEMINI_3.0_FLASH
+ # - GEMINI_API_KEY=<your_gemini_key>
+ #
+ # Azure OpenAI:
# - ENABLE_AZURE=true
- # - LLM_KEY=AZURE_OPENAI # Leave this value static, don't change it
- # - AZURE_DEPLOYMENT=<your_azure_deployment> # Use the OpenAI model "Deployment Name" that you deployed, using the steps above
- # - AZURE_API_KEY=<your_azure_api_key> # Copy and paste Key1 or Key2 from the OpenAI resource in Azure Portal
- # - AZURE_API_BASE=<your_azure_api_base> # Copy and paste the "Endpoint" from the OpenAI resource in Azure Portal (eg. https://xyzxyzxyz.openai.azure.com/)
- # - AZURE_API_VERSION=<your_azure_api_version> # Specify a valid Azure OpenAI data-plane API version (eg. 2024-08-01-preview) Docs: https://learn.microsoft.com/en-us/azure/ai-services/openai/reference
- # Amazon Bedrock Support:
- # Amazon Bedrock is a managed service that enables you to invoke LLMs and bill them through your AWS account.
- # To use Amazon Bedrock as the LLM provider for Skyvern, specify the following environment variables.
- # 1. In the AWS IAM console, create a new AWS IAM User (name it whatever you want)
- # 2. Assign the "AmazonBedrockFullAccess" policy to the user
- # 3. Generate an IAM Access Key under the IAM User's Security Credentials tab
- # 4. In the Amazon Bedrock console, go to "Model Access"
- # 5. Click Modify Model Access button
- # 6. Enable "Claude 3.5 Sonnet v2" and save changes
+ # - LLM_KEY=AZURE_OPENAI
+ # - AZURE_DEPLOYMENT=<your_deployment>
+ # - AZURE_API_KEY=<your_key>
+ # - AZURE_API_BASE=<your_endpoint>
+ # - AZURE_API_VERSION=2024-08-01-preview
+ #
+ # AWS Bedrock:
# - ENABLE_BEDROCK=true
- # - LLM_KEY=BEDROCK_ANTHROPIC_CLAUDE3.5_SONNET # This is the Claude 3.5 Sonnet "V2" model. Change to BEDROCK_ANTHROPIC_CLAUDE3.5_SONNET_V1 for the non-v2 version.
- # - AWS_REGION=us-west-2 # Replace this with a different AWS region, if you desire
- # - AWS_ACCESS_KEY_ID=FILL_ME_IN_PLEASE
- # - AWS_SECRET_ACCESS_KEY=FILL_ME_IN_PLEASE
- # Ollama Support:
- # Ollama is a local LLM provider that can be used to run models locally on your machine.
- # - LLM_KEY=OLLAMA
+ # - LLM_KEY=BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE
+ # - AWS_REGION=us-west-2
+ # - AWS_ACCESS_KEY_ID=<your_key>
+ # - AWS_SECRET_ACCESS_KEY=<your_secret>
+ #
+ # Ollama (local models):
# - ENABLE_OLLAMA=true
- # - OLLAMA_MODEL=qwen2.5:7b-instruct
+ # - LLM_KEY=OLLAMA
+ # - OLLAMA_MODEL=gemma4:e4b
# - OLLAMA_SERVER_URL=http://host.docker.internal:11434
# - OLLAMA_SUPPORTS_VISION=false
- # Open Router Support:
+ #
+ # OpenRouter:
# - ENABLE_OPENROUTER=true
# - LLM_KEY=OPENROUTER
- # - OPENROUTER_API_KEY=<your_openrouter_api_key>
+ # - OPENROUTER_API_KEY=<your_key>
# - OPENROUTER_MODEL=mistralai/mistral-small-3.1-24b-instruct
- # Groq Support:
+ #
+ # Groq:
# - ENABLE_GROQ=true
# - LLM_KEY=GROQ
- # - GROQ_API_KEY=<your_groq_api_key>
- # - GROQ_MODEL=llama-3.1-8b-instant
-
- # Maximum tokens to use: (only set for OpenRouter and Ollama)
- # - LLM_CONFIG_MAX_TOKENS=128000
+ # - GROQ_API_KEY=<your_key>
+ # - GROQ_MODEL=openai/gpt-oss-120b
# Bitwarden Settings
# If you are looking to integrate Skyvern with a password manager (eg Bitwarden), you can use the following environment variables.
diff --git a/docs/developers/debugging/faq.mdx b/docs/developers/debugging/faq.mdx
index 611ce0444..33c996e2b 100644
--- a/docs/developers/debugging/faq.mdx
+++ b/docs/developers/debugging/faq.mdx
@@ -154,7 +154,7 @@ For local installations, run `skyvern init` to generate a local API key, which w
**Cloud:** The model is managed by Skyvern. Contact support for custom model requirements.
**Self-hosted:** Run `skyvern init llm` to reconfigure, or set these environment variables:
-- `LLM_KEY`: Primary model (e.g., `OPENAI_GPT4O`, `ANTHROPIC_CLAUDE_SONNET`)
+- `LLM_KEY`: Primary model (e.g., `OPENAI_GPT5_5`, `ANTHROPIC_CLAUDE4.7_OPUS`)
- `SECONDARY_LLM_KEY`: Optional lightweight model for faster operations
</Accordion>
@@ -352,7 +352,7 @@ Set these environment variables:
```bash
ENABLE_OPENAI=true # or ENABLE_ANTHROPIC, ENABLE_AZURE, etc.
OPENAI_API_KEY=sk-...
-LLM_KEY=OPENAI_GPT4O
+LLM_KEY=OPENAI_GPT5_5
```
Skyvern supports OpenAI, Anthropic, Azure OpenAI, AWS Bedrock, Gemini, and Ollama. Run `skyvern init llm` for guided configuration.
diff --git a/docs/developers/getting-started/quickstart.mdx b/docs/developers/getting-started/quickstart.mdx
index aa798fd98..cd96a1da6 100644
--- a/docs/developers/getting-started/quickstart.mdx
+++ b/docs/developers/getting-started/quickstart.mdx
@@ -236,7 +236,7 @@ ENV='local'
ENABLE_OPENAI='true'
OPENAI_API_KEY='<API_KEY>'
...
-LLM_KEY='OPENAI_GPT4O'
+LLM_KEY='OPENAI_GPT5_5'
SECONDARY_LLM_KEY=''
BROWSER_TYPE='chromium-headful'
MAX_SCRAPING_RETRIES='0'
diff --git a/docs/developers/self-hosted/docker.mdx b/docs/developers/self-hosted/docker.mdx
index 86db8142c..d0c60e9f0 100644
--- a/docs/developers/self-hosted/docker.mdx
+++ b/docs/developers/self-hosted/docker.mdx
@@ -43,7 +43,7 @@ Open `.env` and set your LLM provider. Here's an example for OpenAI:
```bash .env
ENABLE_OPENAI=true
OPENAI_API_KEY=sk-your-api-key-here
-LLM_KEY=OPENAI_GPT4O
+LLM_KEY=OPENAI_GPT5_5
```
For other providers, see [LLM Configuration](/developers/self-hosted/llm-configuration).
@@ -178,12 +178,12 @@ Configure at least one LLM provider. See [LLM Configuration](/developers/self-ho
# OpenAI
ENABLE_OPENAI=true
OPENAI_API_KEY=sk-...
-LLM_KEY=OPENAI_GPT4O
+LLM_KEY=OPENAI_GPT5_5
# Or Anthropic
ENABLE_ANTHROPIC=true
ANTHROPIC_API_KEY=sk-ant-...
-LLM_KEY=ANTHROPIC_CLAUDE3.5_SONNET
+LLM_KEY=ANTHROPIC_CLAUDE4.7_OPUS
```
### Browser settings
diff --git a/docs/developers/self-hosted/kubernetes.mdx b/docs/developers/self-hosted/kubernetes.mdx
index 0084c2cb8..016e17b0b 100644
--- a/docs/developers/self-hosted/kubernetes.mdx
+++ b/docs/developers/self-hosted/kubernetes.mdx
@@ -79,7 +79,7 @@ stringData:
# LLM Configuration - set your provider
ENABLE_OPENAI: "true"
OPENAI_API_KEY: "sk-your-api-key-here"
- LLM_KEY: "OPENAI_GPT4O"
+ LLM_KEY: "OPENAI_GPT5_5"
# Database - points to the PostgreSQL service
DATABASE_STRING: "postgresql+psycopg://skyvern:skyvern@postgres/skyvern"
diff --git a/docs/developers/self-hosted/llm-configuration.mdx b/docs/developers/self-hosted/llm-configuration.mdx
index a68374898..97ece43da 100644
--- a/docs/developers/self-hosted/llm-configuration.mdx
+++ b/docs/developers/self-hosted/llm-configuration.mdx
@@ -32,18 +32,14 @@ For most deployments, configure a single provider using `LLM_KEY`. Skyvern also
## Quick Start Recommendations
-**Best models for production (2025):**
+**Best models for production (2026):**
| Provider | Primary Model | Secondary Model | Notes |
|----------|--------------|-----------------|-------|
-| **Anthropic** | `ANTHROPIC_CLAUDE4.5_OPUS` | `ANTHROPIC_CLAUDE4.5_SONNET` | Most capable |
-| **OpenAI** | `OPENAI_GPT5` | `OPENAI_GPT5_MINI` | Latest |
+| **Anthropic** | `ANTHROPIC_CLAUDE4.7_OPUS` | `ANTHROPIC_CLAUDE4.6_SONNET` | Most capable |
+| **OpenAI** | `OPENAI_GPT5_5` | `OPENAI_GPT5_4` | Latest |
| **Google** | `GEMINI_3_PRO` | `GEMINI_3.0_FLASH` | Latest |
-| **AWS Bedrock** | `BEDROCK_ANTHROPIC_CLAUDE4.5_OPUS_INFERENCE_PROFILE` | `BEDROCK_ANTHROPIC_CLAUDE4.5_SONNET_INFERENCE_PROFILE` | Latest Claude |
-
-<Tip>
-**New in 2025:** GPT-5 series, Claude 4.6 Opus, Gemini 3, Amazon Nova, and many new open-source models via Novita and VolcEngine.
-</Tip>
+| **AWS Bedrock** | `BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE` | `BEDROCK_ANTHROPIC_CLAUDE4.6_SONNET_INFERENCE_PROFILE` | Latest Claude |
---
@@ -54,21 +50,23 @@ The most common choice. Requires an API key from [platform.openai.com](https://p
```bash .env
ENABLE_OPENAI=true
OPENAI_API_KEY=sk-...
-LLM_KEY=OPENAI_GPT4O
+LLM_KEY=OPENAI_GPT5_5
```
### Available models
| LLM_KEY | Notes |
|---------|-------|
+| **GPT-5.5** | |
+| `OPENAI_GPT5_5` | Latest flagship, recommended for complex tasks |
| **GPT-5 Series** | |
-| `OPENAI_GPT5` | Recommended for most complex tasks |
+| `OPENAI_GPT5_4` | |
+| `OPENAI_GPT5` | |
| `OPENAI_GPT5_MINI` | |
| `OPENAI_GPT5_MINI_FLEX` | Flex service tier, 15min timeout |
| `OPENAI_GPT5_NANO` | |
| `OPENAI_GPT5_1` | |
| `OPENAI_GPT5_2` | |
-| `OPENAI_GPT5_4` | |
| **GPT-4 Series** | |
| `OPENAI_GPT4O` | |
| `OPENAI_GPT4O_MINI` | |
@@ -102,18 +100,21 @@ Claude models from [anthropic.com](https://www.anthropic.com/).
```bash .env
ENABLE_ANTHROPIC=true
ANTHROPIC_API_KEY=sk-ant-...
-LLM_KEY=ANTHROPIC_CLAUDE3.5_SONNET
+LLM_KEY=ANTHROPIC_CLAUDE4.7_OPUS
```
### Available models
| LLM_KEY | Notes |
|---------|-------|
+| **Claude 4.7** | |
+| `ANTHROPIC_CLAUDE4.7_OPUS` | Latest, most capable |
| **Claude 4.6** | |
-| `ANTHROPIC_CLAUDE4.6_OPUS` | Newest |
+| `ANTHROPIC_CLAUDE4.6_OPUS` | |
+| `ANTHROPIC_CLAUDE4.6_SONNET` | Recommended for secondary use |
| **Claude 4.5** | |
-| `ANTHROPIC_CLAUDE4.5_OPUS` | Recommended for primary use |
-| `ANTHROPIC_CLAUDE4.5_SONNET` | Recommended for secondary use |
+| `ANTHROPIC_CLAUDE4.5_OPUS` | |
+| `ANTHROPIC_CLAUDE4.5_SONNET` | |
| `ANTHROPIC_CLAUDE4.5_HAIKU` | Fastest |
| **Claude 4** | |
| `ANTHROPIC_CLAUDE4_OPUS` | |
@@ -179,6 +180,7 @@ LLM_KEY=GEMINI_2.5_PRO
| **Gemini 3** | |
| `GEMINI_3_PRO` | Recommended for primary use |
| `GEMINI_3.0_FLASH` | Recommended for secondary use |
+| `GEMINI_3.1_FLASH_LITE` | Fastest, most cost-efficient |
| **Gemini 2.5** | |
| `GEMINI_2.5_PRO` | |
| `GEMINI_2.5_PRO_PREVIEW` | |
@@ -227,6 +229,7 @@ For global endpoint access, set `VERTEX_LOCATION=global` and ensure `VERTEX_PROJ
| **Gemini 3** | |
| `VERTEX_GEMINI_3_PRO` | Recommended for primary use |
| `VERTEX_GEMINI_3.0_FLASH` | Recommended for secondary use |
+| `VERTEX_GEMINI_3.1_FLASH_LITE` | Fastest, most cost-efficient |
| **Gemini 2.5** | |
| `VERTEX_GEMINI_2.5_PRO` | |
| `VERTEX_GEMINI_2.5_PRO_PREVIEW` | |
@@ -247,7 +250,7 @@ Run Anthropic Claude through your AWS account.
```bash .env
ENABLE_BEDROCK=true
-LLM_KEY=BEDROCK_ANTHROPIC_CLAUDE3.5_SONNET
+LLM_KEY=BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE
AWS_REGION=us-west-2
AWS_ACCESS_KEY_ID=AKIA...
AWS_SECRET_ACCESS_KEY=...
@@ -258,7 +261,7 @@ AWS_SECRET_ACCESS_KEY=...
1. Create an IAM user with `AmazonBedrockFullAccess` policy
2. Generate access keys for the IAM user
3. In the [Bedrock console](https://console.aws.amazon.com/bedrock/), go to **Model Access**
-4. Enable access to Claude 3.5 Sonnet
+4. Enable access to Claude Opus 4.7
### Available models
@@ -267,8 +270,11 @@ AWS_SECRET_ACCESS_KEY=...
| **Amazon Nova (AWS Native)** | |
| `BEDROCK_AMAZON_NOVA_PRO` | |
| `BEDROCK_AMAZON_NOVA_LITE` | |
+| **Claude 4.7** | |
+| `BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE` | Cross-region, latest |
| **Claude 4.6** | |
| `BEDROCK_ANTHROPIC_CLAUDE4.6_OPUS_INFERENCE_PROFILE` | Cross-region |
+| `BEDROCK_ANTHROPIC_CLAUDE4.6_SONNET_INFERENCE_PROFILE` | Cross-region |
| **Claude 4.5** | |
| `BEDROCK_ANTHROPIC_CLAUDE4.5_OPUS_INFERENCE_PROFILE` | Cross-region |
| `BEDROCK_ANTHROPIC_CLAUDE4.5_SONNET_INFERENCE_PROFILE` | Cross-region |
@@ -436,7 +442,7 @@ Run open-source models locally with [Ollama](https://ollama.ai/). No API costs,
```bash .env
ENABLE_OLLAMA=true
LLM_KEY=OLLAMA
-OLLAMA_MODEL=llama3.1
+OLLAMA_MODEL=gemma4
OLLAMA_SERVER_URL=http://host.docker.internal:11434
OLLAMA_SUPPORTS_VISION=false
```
@@ -444,7 +450,7 @@ OLLAMA_SUPPORTS_VISION=false
### Setup steps
1. [Install Ollama](https://ollama.ai/download)
-2. Pull a model: `ollama pull llama3.1`
+2. Pull a model: `ollama pull gemma4`
3. Start Ollama: `ollama serve`
4. Configure Skyvern to connect
@@ -469,7 +475,7 @@ Connect to any service that implements the OpenAI API format, including LiteLLM,
```bash .env
ENABLE_OPENAI_COMPATIBLE=true
-OPENAI_COMPATIBLE_MODEL_NAME=llama3.1
+OPENAI_COMPATIBLE_MODEL_NAME=gemma4
OPENAI_COMPATIBLE_API_KEY=sk-test
OPENAI_COMPATIBLE_API_BASE=http://localhost:4000/v1
LLM_KEY=OPENAI_COMPATIBLE
@@ -520,25 +526,25 @@ Configure a cheaper model for lightweight operations:
```bash .env
# Main model for complex decisions
-LLM_KEY=ANTHROPIC_CLAUDE4.5_OPUS
-# or: OPENAI_GPT5
+LLM_KEY=ANTHROPIC_CLAUDE4.7_OPUS
+# or: OPENAI_GPT5_5
# or: GEMINI_3_PRO
# Faster model for simple tasks like dropdown selection
-SECONDARY_LLM_KEY=ANTHROPIC_CLAUDE4.5_SONNET
-# or: OPENAI_GPT5_MINI
+SECONDARY_LLM_KEY=ANTHROPIC_CLAUDE4.6_SONNET
+# or: OPENAI_GPT5_4
# or: GEMINI_3.0_FLASH
```
<Tip>
**Recommended primary models (latest):**
-- **Anthropic Claude 4.5 Opus** (`ANTHROPIC_CLAUDE4.5_OPUS`) - Most capable
-- **OpenAI GPT-5** (`OPENAI_GPT5`) - Latest
+- **Anthropic Claude 4.7 Opus** (`ANTHROPIC_CLAUDE4.7_OPUS`) - Most capable
+- **OpenAI GPT-5.5** (`OPENAI_GPT5_5`) - Latest
- **Google Gemini 3 Pro** (`GEMINI_3_PRO`) - Latest
**Recommended secondary models (latest):**
-- **Claude 4.5 Sonnet** (`ANTHROPIC_CLAUDE4.5_SONNET`) - Balanced
-- **GPT-5 Mini** (`OPENAI_GPT5_MINI`) - Faster GPT-5
+- **Claude 4.6 Sonnet** (`ANTHROPIC_CLAUDE4.6_SONNET`) - Balanced
+- **GPT-5.4** (`OPENAI_GPT5_4`) - Fast reasoning
- **Gemini 3.0 Flash** (`GEMINI_3.0_FLASH`) - Faster Gemini 3
</Tip>
@@ -565,13 +571,13 @@ Most deployments don't need task-specific models. Start with `LLM_KEY` and `SECO
Some operations require a secondary model. Set `SECONDARY_LLM_KEY` in your environment:
```bash .env
-SECONDARY_LLM_KEY=OPENAI_GPT4O_MINI
+SECONDARY_LLM_KEY=OPENAI_GPT5_4
```
### "Context window exceeded"
The page content is too large for the model's context window. Options:
-- Use a model with larger context support (GPT-5, Gemini 2.5 Pro, or Claude 4.5 Sonnet)
+- Use a model with larger context support (GPT-5.5, Gemini 2.5 Pro, or Claude 4.7 Opus)
- Simplify your prompt to require less page analysis
- Start from a more specific URL with less content
@@ -580,7 +586,7 @@ The page content is too large for the model's context window. Options:
The configured `LLM_KEY` doesn't match any enabled provider. Verify:
1. The provider is enabled (`ENABLE_OPENAI=true`, etc.)
2. The `LLM_KEY` value matches a supported model name exactly
-3. Model names are case-sensitive: `OPENAI_GPT4O` not `openai_gpt4o`
+3. Model names are case-sensitive: `OPENAI_GPT5_5` not `openai_gpt5_5`
### Container logs show authentication errors
diff --git a/docs/integrations/local-llms.mdx b/docs/integrations/local-llms.mdx
index 63a8ec0a9..e9513d9ca 100644
--- a/docs/integrations/local-llms.mdx
+++ b/docs/integrations/local-llms.mdx
@@ -31,7 +31,7 @@ Use Ollama to run open-source models locally.
### Step 1: Start Ollama
```bash
-ollama pull llama3.1
+ollama pull gemma4
ollama serve
```
@@ -44,7 +44,7 @@ Add to your `.env` file:
```bash
ENABLE_OLLAMA=true
OLLAMA_SERVER_URL=http://localhost:11434
-OLLAMA_MODEL=llama3.1
+OLLAMA_MODEL=gemma4
# Enable for vision models (qwen2-vl, llava, etc.)
OLLAMA_SUPPORTS_VISION=false
@@ -72,7 +72,7 @@ Use LiteLLM as an OpenAI-compatible proxy to connect any model provider.
### Step 1: Start LiteLLM
```bash
-litellm --model ollama/llama3.1 --host 0.0.0.0 --port 4000
+litellm --model ollama/gemma4 --host 0.0.0.0 --port 4000
```
### Step 2: Configure Skyvern
@@ -81,7 +81,7 @@ Add to your `.env` file:
```bash
ENABLE_OPENAI_COMPATIBLE=true
-OPENAI_COMPATIBLE_MODEL_NAME=llama3.1
+OPENAI_COMPATIBLE_MODEL_NAME=gemma4
OPENAI_COMPATIBLE_API_KEY=sk-test
OPENAI_COMPATIBLE_API_BASE=http://localhost:4000/v1
```
diff --git a/skyvern/analytics.py b/skyvern/analytics.py
index 6d8695585..51d348ac0 100644
--- a/skyvern/analytics.py
+++ b/skyvern/analytics.py
@@ -1,6 +1,8 @@
+import atexit
import functools
import importlib.metadata
import platform
+import threading
import traceback
from typing import Any, Dict, Optional
@@ -13,12 +15,19 @@ from skyvern.config import settings
LOG = structlog.get_logger(__name__)
+_FLUSH_TIMEOUT_SECONDS = 2
-def _build_posthog_client(api_key: str, host: str) -> Posthog:
- return Posthog(api_key, host=host, disable_geoip=False, timeout=2)
+def _build_posthog_client(api_key: str, host: str) -> Posthog | None:
+ try:
+ client = Posthog(api_key, host=host, disable_geoip=False, timeout=2, max_retries=1)
+ atexit.unregister(client.join)
+ return client
+ except Exception:
+ return None
-posthog = _build_posthog_client(
+
+posthog: Posthog | None = _build_posthog_client(
settings.POSTHOG_PROJECT_API_KEY,
settings.POSTHOG_PROJECT_HOST,
)
@@ -72,7 +81,7 @@ def reconfigure_posthog_client(
def _resolve_posthog_client(
api_key: str | None = None,
host: str | None = None,
-) -> Posthog:
+) -> Posthog | None:
if api_key is None and host is None:
return posthog
@@ -82,7 +91,8 @@ def _resolve_posthog_client(
client = _custom_posthog_clients.get(cache_key)
if client is None:
client = _build_posthog_client(resolved_api_key, resolved_host)
- _custom_posthog_clients[cache_key] = client
+ if client is not None:
+ _custom_posthog_clients[cache_key] = client
return client
@@ -90,7 +100,12 @@ def flush(
api_key: str | None = None,
host: str | None = None,
) -> None:
- _resolve_posthog_client(api_key=api_key, host=host).flush()
+ client = _resolve_posthog_client(api_key=api_key, host=host)
+ if client is None:
+ return
+ t = threading.Thread(target=client.flush, daemon=True)
+ t.start()
+ t.join(timeout=_FLUSH_TIMEOUT_SECONDS)
def capture(
@@ -104,9 +119,11 @@ def capture(
return
try:
+ client = _resolve_posthog_client(api_key=api_key, host=host)
+ if client is None:
+ return
resolved_distinct_id = distinct_id or settings.ANALYTICS_ID
payload: dict[str, Any] = {**dynamic_analytics_metadata(), **(data or {})}
- client = _resolve_posthog_client(api_key=api_key, host=host)
client.capture(distinct_id=resolved_distinct_id, event=event, properties=payload)
except Exception:
LOG.debug("analytics capture failed", event=event, exc_info=True)
diff --git a/skyvern/cli/llm_setup.py b/skyvern/cli/llm_setup.py
index ac2112a94..be1027b17 100644
--- a/skyvern/cli/llm_setup.py
+++ b/skyvern/cli/llm_setup.py
@@ -76,13 +76,9 @@ def setup_llm_providers() -> None:
enabled_providers.append("openai")
model_options.extend(
[
+ "OPENAI_GPT5_5",
"OPENAI_GPT5_4",
"OPENAI_GPT5",
- "OPENAI_GPT5_2",
- "OPENAI_GPT4_1",
- "OPENAI_O3",
- "OPENAI_O4_MINI",
- "OPENAI_GPT4O",
]
)
else:
@@ -101,8 +97,9 @@ def setup_llm_providers() -> None:
enabled_providers.append("anthropic")
model_options.extend(
[
+ "ANTHROPIC_CLAUDE4.7_OPUS",
"ANTHROPIC_CLAUDE4.6_OPUS",
- "ANTHROPIC_CLAUDE4.5_OPUS",
+ "ANTHROPIC_CLAUDE4.6_SONNET",
"ANTHROPIC_CLAUDE4.5_SONNET",
"ANTHROPIC_CLAUDE4.5_HAIKU",
]
@@ -142,11 +139,21 @@ def setup_llm_providers() -> None:
update_or_add_env_var("GEMINI_API_KEY", gemini_api_key)
update_or_add_env_var("ENABLE_GEMINI", "true")
enabled_providers.append("gemini")
+ model_options.extend(
+ [
+ "GEMINI_3.1_PRO",
+ "GEMINI_3.0_FLASH",
+ "GEMINI_3.1_FLASH_LITE",
+ "GEMINI_2.5_PRO",
+ "GEMINI_2.5_FLASH",
+ "GEMINI_2.5_FLASH_LITE",
+ ]
+ )
else:
update_or_add_env_var("ENABLE_GEMINI", "false")
console.print("\n[bold blue]--- Ollama / Local LLM Configuration ---[/bold blue]")
- console.print("Use any locally-running model via Ollama (e.g. llama3, mistral, qwen).")
+ console.print("Use any locally-running model via Ollama (e.g. gemma4, qwen3, deepseek-r1).")
console.print("[dim]Requires Ollama running locally: https://ollama.com[/dim]")
enable_ollama = Confirm.ask("Do you want to enable a local Ollama model?")
if enable_ollama:
@@ -155,7 +162,7 @@ def setup_llm_providers() -> None:
default="http://localhost:11434",
)
ollama_model = Prompt.ask(
- "Enter model name (e.g. llama3.3, mistral, qwen2.5)",
+ "Enter model name (e.g. gemma4, qwen3, deepseek-r1)",
)
if not ollama_model:
console.print("[red]Error: Model name is required. Ollama will not be enabled.[/red]")
diff --git a/skyvern/cli/quickstart.py b/skyvern/cli/quickstart.py
index 4b4d4e321..52482838a 100644
--- a/skyvern/cli/quickstart.py
+++ b/skyvern/cli/quickstart.py
@@ -12,6 +12,7 @@ from rich.prompt import Confirm
from skyvern.analytics import capture_setup_error, capture_setup_event
# Import console after skyvern.cli to ensure proper initialization
+from skyvern.cli.browser import _open_chrome_inspect
from skyvern.cli.console import console
from skyvern.cli.init_command import init_env # init is used directly
from skyvern.cli.llm_setup import setup_llm_providers
@@ -142,6 +143,37 @@ def run_docker_compose_setup() -> None:
)
)
+ # Offer to set up "Control your own browser"
+ use_own_browser = Confirm.ask(
+ "\nWould you like to [bold yellow]control your own Chrome browser[/bold yellow] (use your cookies, logins, and extensions)?",
+ default=False,
+ )
+ if use_own_browser:
+ console.print(
+ Panel(
+ "[bold]Enable Remote Debugging in Chrome[/bold]\n\n"
+ "1. We'll open [cyan]chrome://inspect/#remote-debugging[/cyan] in your browser\n"
+ "2. Click [bold]Enable[/bold] to start the debugging server\n"
+ "3. You should see: [green]Server running at: 127.0.0.1:9222[/green]",
+ border_style="cyan",
+ )
+ )
+ open_page = Confirm.ask("Open chrome://inspect/#remote-debugging now?", default=True)
+ if open_page:
+ _open_chrome_inspect()
+ confirmed = Confirm.ask("Have you enabled remote debugging in Chrome?", default=False)
+ if confirmed:
+ from skyvern.cli.llm_setup import update_or_add_env_var
+
+ update_or_add_env_var("BROWSER_TYPE", "cdp-connect")
+ update_or_add_env_var("BROWSER_REMOTE_DEBUGGING_URL", "http://host.docker.internal:9222/")
+ console.print("✅ [green]Browser debugging configured in .env. Restart with:[/green]")
+ console.print(" [cyan]docker compose up -d[/cyan]")
+ else:
+ console.print(
+ "[yellow]No problem — you can enable it later by navigating to chrome://inspect/#remote-debugging in Chrome.[/yellow]"
+ )
+
@quickstart_app.callback(invoke_without_command=True)
def quickstart(
diff --git a/skyvern/config.py b/skyvern/config.py
index 82adf8642..c784961f9 100644
--- a/skyvern/config.py
+++ b/skyvern/config.py
@@ -213,7 +213,7 @@ class Settings(BaseSettings):
# LLM Configuration #
#####################
# ACTIVE LLM PROVIDER
- LLM_KEY: str = "OPENAI_GPT4O" # This is the model name
+ LLM_KEY: str = "OPENAI_GPT5_5" # This is the model name
LLM_API_KEY: str | None = None # API key for the model
SECONDARY_LLM_KEY: str | None = None
SELECT_AGENT_LLM_KEY: str | None = None
@@ -252,7 +252,7 @@ class Settings(BaseSettings):
GPT5_REASONING_EFFORT: str | None = "medium"
# ANTHROPIC
ANTHROPIC_API_KEY: str | None = None
- ANTHROPIC_CUA_LLM_KEY: str = "ANTHROPIC_CLAUDE3.7_SONNET"
+ ANTHROPIC_CUA_LLM_KEY: str = "ANTHROPIC_CLAUDE4.6_SONNET"
# VOLCENGINE (Doubao)
ENABLE_VOLCENGINE: bool = False
diff --git a/skyvern/forge/sdk/api/llm/config_registry.py b/skyvern/forge/sdk/api/llm/config_registry.py
index 967791346..7b17436ae 100644
--- a/skyvern/forge/sdk/api/llm/config_registry.py
+++ b/skyvern/forge/sdk/api/llm/config_registry.py
@@ -215,6 +215,18 @@ if settings.ENABLE_OPENAI:
),
),
)
+ LLMConfigRegistry.register_config(
+ "OPENAI_GPT5_5",
+ LLMConfig(
+ "gpt-5.5",
+ ["OPENAI_API_KEY"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=128000,
+ temperature=1, # GPT-5 only supports temperature=1
+ reasoning_effort=settings.GPT5_REASONING_EFFORT,
+ ),
+ )
LLMConfigRegistry.register_config(
"OPENAI_GPT4_TURBO",
LLMConfig(
@@ -471,7 +483,28 @@ if settings.ENABLE_ANTHROPIC:
temperature=1, # Claude 4.6 only supports temperature=1
),
)
-
+ LLMConfigRegistry.register_config(
+ "ANTHROPIC_CLAUDE4.6_SONNET",
+ LLMConfig(
+ "anthropic/claude-sonnet-4-6",
+ ["ANTHROPIC_API_KEY"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=64000,
+ temperature=1,
+ ),
+ )
+ LLMConfigRegistry.register_config(
+ "ANTHROPIC_CLAUDE4.7_OPUS",
+ LLMConfig(
+ "anthropic/claude-opus-4-7",
+ ["ANTHROPIC_API_KEY"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=128000,
+ temperature=1,
+ ),
+ )
if settings.ENABLE_BEDROCK:
# Supported through AWS IAM authentication
LLMConfigRegistry.register_config(
@@ -623,6 +656,28 @@ if settings.ENABLE_BEDROCK:
temperature=1, # Claude 4.6 only supports temperature=1
),
)
+ LLMConfigRegistry.register_config(
+ "BEDROCK_ANTHROPIC_CLAUDE4.6_SONNET_INFERENCE_PROFILE",
+ LLMConfig(
+ "bedrock/us.anthropic.claude-sonnet-4-6",
+ ["AWS_REGION"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=64000,
+ temperature=1,
+ ),
+ )
+ LLMConfigRegistry.register_config(
+ "BEDROCK_ANTHROPIC_CLAUDE4.7_OPUS_INFERENCE_PROFILE",
+ LLMConfig(
+ "bedrock/us.anthropic.claude-opus-4-7",
+ ["AWS_REGION"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=128000,
+ temperature=1,
+ ),
+ )
if settings.ENABLE_AZURE:
@@ -1217,6 +1272,19 @@ if settings.ENABLE_GEMINI:
),
),
)
+ LLMConfigRegistry.register_config(
+ "GEMINI_3.1_FLASH_LITE",
+ LLMConfig(
+ "gemini/gemini-3.1-flash-lite-preview",
+ ["GEMINI_API_KEY"],
+ supports_vision=True,
+ add_assistant_prefix=False,
+ max_completion_tokens=65536,
+ litellm_params=LiteLLMParams(
+ thinking_level="medium" if settings.GEMINI_INCLUDE_THOUGHT else "minimal",
+ ),
+ ),
+ )
# Backward compat alias for non-Vertex Gemini 3 Pro
LLMConfigRegistry.register_config(
"GEMINI_3.1_PRO",
diff --git a/tests/sdk/README.md b/tests/sdk/README.md
index e91c341ef..60e053c96 100644
--- a/tests/sdk/README.md
+++ b/tests/sdk/README.md
@@ -36,17 +36,9 @@ pytest tests/sdk/python_sdk/test_sdk_simple_actions.py::test_clicks
- Requires Chrome/Chromium with CDP on `localhost:9222` (see below)
- Web server auto-starts via `run-test.js`
-**Launch Chromium with CDP:**
-```bash
-# Find your Playwright Chromium path
-ls ~/Library/Caches/ms-playwright/
-
-# Then launch it (adjust the chromium-XXXX version to match yours)
-~/Library/Caches/ms-playwright/chromium-XXXX/chrome-mac/Chromium.app/Contents/MacOS/Chromium \
- --remote-debugging-port=9222 \
- --user-data-dir=~/tmp/chrome-playwright \
- about:blank
-```
+**Enable Chrome remote debugging:**
+1. Open Chrome and navigate to `chrome://inspect/#remote-debugging`
+2. Click **Enable** — you should see "Server running at: 127.0.0.1:9222"
### Running Tests