docs: convert screenshots to webp and update documentation
|
After Width: | Height: | Size: 39 KiB |
|
After Width: | Height: | Size: 93 KiB |
@@ -61,12 +61,12 @@ Configures whether background transcoding and AI agents are executed locally or
|
||||
|
||||
---
|
||||
|
||||
## AI Credentials Backup
|
||||
## LLM API Key
|
||||
|
||||
Optional global keys to pre-authorize the built-in AI providers on startup.
|
||||
API keys to AI providers on startup.
|
||||
|
||||
| Variable | Description | Example / Default |
|
||||
| :------------------ | :------------------------------------- | :---------------- |
|
||||
| `GEMINI_API_KEY` | Global key for Google Gemini provider. | `AIzaSy...` |
|
||||
| `OPENAI_API_KEY` | Global key for OpenAI provider. | `sk-proj-...` |
|
||||
| `ANTHROPIC_API_KEY` | Global key for Anthropic provider. | `sk-ant-...` |
|
||||
| Variable | Description | Example / Default |
|
||||
| :------------------ | :---------------------------------- | :---------------- |
|
||||
| `GEMINI_API_KEY` | API key for Google Gemini provider. | `AIzaSy...` |
|
||||
| `OPENAI_API_KEY` | API key for OpenAI provider. | `sk-proj-...` |
|
||||
| `ANTHROPIC_API_KEY` | API key for Anthropic provider. | `sk-ant-...` |
|
||||
|
||||
@@ -25,6 +25,7 @@
|
||||
"using-shumai/comments",
|
||||
"using-shumai/version-stack",
|
||||
"using-shumai/sort-and-search",
|
||||
"using-shumai/semantic-search",
|
||||
"using-shumai/team-project-management",
|
||||
"using-shumai/notifications",
|
||||
"using-shumai/metadata",
|
||||
@@ -33,7 +34,7 @@
|
||||
]
|
||||
},
|
||||
{
|
||||
"group": "Shumai Agent (Admin)",
|
||||
"group": "Shumai Agent Settings",
|
||||
"pages": [
|
||||
"using-shumai/agent-providers",
|
||||
"using-shumai/agents",
|
||||
@@ -42,20 +43,16 @@
|
||||
]
|
||||
},
|
||||
{
|
||||
"group": "Media Transcoding (Admin)",
|
||||
"group": "Media Transcoding Settings",
|
||||
"pages": ["using-shumai/transcode-settings"]
|
||||
},
|
||||
{
|
||||
"group": "Advanced Search",
|
||||
"pages": ["using-shumai/semantic-search"]
|
||||
},
|
||||
{
|
||||
"group": "Advanced Deployment",
|
||||
"pages": ["configuration/temporal"]
|
||||
},
|
||||
{
|
||||
"group": "Configuration",
|
||||
"pages": ["configuration/env-variables"]
|
||||
},
|
||||
{
|
||||
"group": "Temporal Deployment",
|
||||
"pages": ["configuration/temporal"]
|
||||
}
|
||||
]
|
||||
}
|
||||
|
||||
|
After Width: | Height: | Size: 87 KiB |
|
After Width: | Height: | Size: 94 KiB |
|
After Width: | Height: | Size: 23 KiB |
|
After Width: | Height: | Size: 51 KiB |
|
After Width: | Height: | Size: 40 KiB |
|
After Width: | Height: | Size: 52 KiB |
|
After Width: | Height: | Size: 65 KiB |
|
After Width: | Height: | Size: 70 KiB |
|
After Width: | Height: | Size: 45 KiB |
|
After Width: | Height: | Size: 39 KiB |
|
After Width: | Height: | Size: 49 KiB |
|
After Width: | Height: | Size: 32 KiB |
@@ -3,36 +3,47 @@ title: AI Providers
|
||||
description: Connect Shumai to OpenAI, Anthropic, Gemini, or custom LLM gateways.
|
||||
---
|
||||
|
||||
To enable AI Agents, metadata autofill, and semantic search, Shumai must be connected to an AI model provider. Team Owners can manage these credentials under the **AI Providers** settings tab.
|
||||
When creating agents for chat or metadata autofill, you must select an underlying AI model. The available models and providers are configured globally within your team settings.
|
||||
|
||||
> 🔒 **Access Control:** Only **Team Owner** can manage AI providers. To update these settings, click your avatar, navigate to **Settings**, and select **Providers** from the left sidebar.
|
||||
|
||||
---
|
||||
|
||||
## Built-In Providers
|
||||
|
||||
Shumai ships with pre-configured settings for major LLM providers:
|
||||
Shumai ships with pre-configured configurations for major LLM providers. Because Shumai uses [Pi](https://github.com/earendil-works/pi) as its core agent framework, the supported provider and model configurations are automatically synchronized with Pi's capabilities.
|
||||
|
||||
- **Google Gemini**: Recommended for multimodal asset tagging (Autofill Agent) and reasoning tasks.
|
||||
- **OpenAI**: Supports GPT-4, GPT-4o, and reasoning models.
|
||||
- **Anthropic**: Supports the Claude 3.5 Sonnet and Haiku models.
|
||||
|
||||
### Global Credentials
|
||||
|
||||
If your system administrator has configured global keys (like `GEMINI_API_KEY`, `OPENAI_API_KEY`, or `ANTHROPIC_API_KEY`) in the server's environment variables, the respective built-in provider will use these credentials automatically, making them immediately available to the entire team.
|
||||
Team Owner can freely edit, delete, or add new providers. All configured provider and model settings are injected directly into the Pi framework at runtime.
|
||||
|
||||
---
|
||||
|
||||
## Custom Providers
|
||||
## Configuration Reference
|
||||
|
||||
If you want to use local models (via Ollama or vLLM) or private enterprise proxy gateways, you can register custom providers:
|
||||
### Provider Settings
|
||||
|
||||
1. Navigate to **Team Settings** > **AI Providers**.
|
||||
2. Click **Add Provider**.
|
||||
3. Fill in the configuration parameters:
|
||||
- **Provider Name**: E.g., `Ollama-Local` or `Enterprise-Gateway`.
|
||||
- **Base URL**: The target API URL (e.g., `http://localhost:11434/v1` for Ollama).
|
||||
- **API Key**: The authentication token (use `nokey` if local/unauthenticated).
|
||||
4. **Register Models**: Define the list of model IDs and names (e.g., model ID `llama3`, name `Llama 3`) that this provider exposes.
|
||||
5. Click **Save**.
|
||||
When configuring a provider, you will need to specify the following fields:
|
||||
|
||||

|
||||
_Placeholder: Table view of active AI providers showing their status, built-in badges, and active models._
|
||||
| Field | Description | Example / Notes |
|
||||
| :---------------- | :------------------------------------------------------------------------- | :------------------------------------------------------------------------------------------------------------------------------- |
|
||||
| **Provider Name** | A unique, human-readable label for the provider. | `Ollama-Local`, `Enterprise-Gateway` |
|
||||
| **Base URL** | The target API endpoint URL. | `http://localhost:11434/v1` |
|
||||
| **API Protocol** | The streaming and payload implementation format. | Most OpenAI-compatible gateways use `openai-completions`. |
|
||||
| **API Key** | The authentication token. Leave blank for unauthenticated or local setups. | Accepts a literal value (`sk-...`) or an environment variable name (`MY_API_KEY`) to load secrets securely from the host system. |
|
||||
|
||||
### Model Settings
|
||||
|
||||
Each provider can host multiple models. You can configure the following attributes per model:
|
||||
|
||||
| Field | Description | Example / Notes |
|
||||
| :-------------------- | :---------------------------------------------------------- | :------------------------------------------------- |
|
||||
| **Model ID** | The exact identifier string expected by the upstream API. | `gpt-4o`, `claude-3-5-sonnet-20240620` |
|
||||
| **Display Name** | An optional, friendly label shown to users in dropdowns. | `GPT-4 Omni`, `Claude 3.5` |
|
||||
| **Reasoning Support** | Toggle to enable extended thinking protocols. | Use for models with native reasoning capabilities. |
|
||||
| **Context Window** | The maximum total token capacity (input + output). | |
|
||||
| **Max Output Tokens** | The hard ceiling for generated response tokens per request. | |
|
||||
| **Input Cost** | Cost tracking specified per million input tokens. | Optional; used to calculate usage metrics. |
|
||||
| **Output Cost** | Cost tracking specified per million output tokens. | Optional; used to calculate usage metrics. |
|
||||
|
||||
---
|
||||
|
||||

|
||||
|
||||
@@ -14,6 +14,13 @@ The sandbox restricts the execution process from accessing the host filesystem,
|
||||
- **Isolation**: Scripts run in a virtualized workspace container that prevents them from modifying critical system files or reading other project data.
|
||||
- **Network Bridging**: External network calls from inside the sandbox are strictly monitored.
|
||||
|
||||
### File System Restrictions
|
||||
|
||||
To protect your workspace environments and sensitive data, the sandbox enforces strict file-level access controls:
|
||||
|
||||
- **Read Access**: The sandbox is allowed to read files within the workspace, with a strict exception for sensitive environment and certificate files. It cannot read: `['.env', '.env.*', '*.pem', '*.key']`.
|
||||
- **Write Access**: The sandbox is heavily restricted and may only write to the `.pi` folder located in the current working directory (cwd). Even within this allowed directory, the sandbox is explicitly blocked from writing or creating sensitive files matching: `['.env', '.env.*', '*.pem', '*.key']`.
|
||||
|
||||
---
|
||||
|
||||
## Domain Whitelisting
|
||||
@@ -35,5 +42,4 @@ When an agent script attempts to reach a domain that is not currently authorized
|
||||
2. The target domain is logged and added to the **Pending Domains** list.
|
||||
3. Admins can review this list and click **Approve** to authorize the domain, or **Reject** to keep it blocked.
|
||||
|
||||

|
||||
_Placeholder: Security settings page showing lists of Allowed Domains, Pending Domains with Approve/Deny buttons, and sandbox audit logs._
|
||||

|
||||
|
||||
@@ -3,7 +3,7 @@ title: Agent Skills
|
||||
description: Extend your agents' capabilities with custom automation tools.
|
||||
---
|
||||
|
||||
By default, AI models can only process text. In Shumai, you can extend your agents' capabilities by registering **Skills**—custom scripts that the agent can run when solving a task.
|
||||
By default, AI models can only process text and image. In Shumai, you can extend both your **Chat Agents** and **Autofill Agents'** capabilities by registering **Skills**—custom scripts that the agent can run when solving a task.
|
||||
|
||||
---
|
||||
|
||||
@@ -11,23 +11,44 @@ By default, AI models can only process text. In Shumai, you can extend your agen
|
||||
|
||||
Skills are actions that you define and make available to your AI agents.
|
||||
|
||||
- **Descriptions for the AI**: When you register a skill, you write a clear description of what the skill does. The AI agent reads this description to understand when it should use the tool.
|
||||
- **Tool Execution**: When the Chat Agent decides it needs to use a skill (for example, to post a message to Slack or look up information), it requests to run that skill. Shumai executes the script in the background and returns the results to the agent.
|
||||
- **Descriptions for the AI**: When you register a skill, the AI agent reads its metadata to understand exactly when and how it should use the tool to solve a problem.
|
||||
- **Execution & Local Caching**: When an agent decides it needs to use a skill, Shumai handles the execution. Newly added skill data is synced to the database first. Upon execution, the agent downloads and caches the skill locally from the database for fast, reliable access.
|
||||
|
||||
---
|
||||
|
||||
## Registering a Skill
|
||||
## Adding a Skill
|
||||
|
||||
To add a new skill, navigate to **Team Settings** > **Skills**:
|
||||
To install a new skill into Shumai:
|
||||
|
||||
1. Click **Create Skill**.
|
||||
2. Fill in the parameters:
|
||||
- **Name**: E.g., `image_resizer` or `slack_notify`.
|
||||
- **Description**: A clear explanation of what the tool does (e.g. "Use this tool to send a Slack message to the team channel").
|
||||
- **Parameters**: Define the input parameters the agent must provide (such as the target channel name and the message text).
|
||||
- **Script Code**: The automation script code uploaded by your administrator.
|
||||
3. Click **Save**.
|
||||
4. Navigate to **Agents** and check the box to assign this skill to your chosen Chat Agent.
|
||||
1. Navigate to the **Settings** page and open the **Skills** tab.
|
||||
2. Click the **Add Skill** button.
|
||||
3. In the dialog box, choose your installation method:
|
||||
- **Upload**: Select and upload a `.zip` file containing your skill.
|
||||
- **GitHub**: Provide a valid GitHub repository path.
|
||||
|
||||

|
||||
_Placeholder: Table displaying custom skills, their descriptions, code editors, and the list of agents currently utilizing them._
|
||||
> **Important:** Regardless of the installation method, the root folder of the skill must contain a `SKILL.md` file.
|
||||
|
||||
### Updating an Existing Skill
|
||||
|
||||
To update a skill, simply follow the steps above and install the new version using the exact same skill name. Shumai will automatically replace the older version.
|
||||
|
||||
---
|
||||
|
||||
## Configuring Environment Variables
|
||||
|
||||
If your custom skill requires environment variables (e.g., API keys, external tokens) to function:
|
||||
|
||||
1. Click on the installed **Skill Card** in the Skills tab to open the configuration dialog.
|
||||
2. Add the required environment variables.
|
||||
3. _(Optional)_ Set a default value for the variable. If no default value is provided, the skill will dynamically load the required value from the system environment at runtime.
|
||||
|
||||
---
|
||||
|
||||
## Enabling Skills for Agents
|
||||
|
||||
**All skills are disabled by default.** Merely installing a skill does not give agents access to it. To allow an agent to use an installed skill:
|
||||
|
||||
1. Navigate to your specific **Agent Settings**.
|
||||
2. Explicitly enable the skills you want to authorize for that particular agent.
|
||||
|
||||

|
||||
|
||||
@@ -3,7 +3,7 @@ title: AI Agents
|
||||
description: Create and configure specialized AI agents for chat, metadata autofill, and vector embedding.
|
||||
---
|
||||
|
||||
Shumai uses specialized AI Agents to perform background automation, chat workflows, and semantic indexing. Team Owners can create and manage multiple agents.
|
||||
Shumai uses specialized AI Agents to perform background automation, chat workflows, and semantic indexing. **Only Team Owner** can create, manage, and change agent settings.
|
||||
|
||||
---
|
||||
|
||||
@@ -15,35 +15,36 @@ Shumai supports three types of agents:
|
||||
|
||||
An interactive, conversational bot that appears in your project workspace.
|
||||
|
||||
- You can start an active chat session with the agent inside any project.
|
||||
- You can create as many chat agents as you want.
|
||||
- You can use "@" to mention chat agents directly in file comments.
|
||||
- The agent is context-aware: it can read files, search the project folder, inspect metadata, and read comment threads.
|
||||
- You can attach custom "Skills" to a chat agent, enabling it to write files, run scripts, or interact with external APIs on your behalf.
|
||||
- You can add "Skills" to a chat agent, enabling it to write files, run scripts, or interact with external APIs on your behalf.
|
||||
|
||||
### 2. Autofill Agent
|
||||
|
||||
A background agent that monitors file uploads.
|
||||
A background agent that monitors file uploads. You can create **only one** Autofill Agent.
|
||||
|
||||
- Once a file is uploaded, the Autofill Agent is triggered.
|
||||
- Once created and enabled, Shumai automatically triggers the agent when a new file is uploaded, provided the file type is supported (image or video).
|
||||
- It analyzes the file's content (multimodal images, text descriptions) and populates any metadata fields flagged for "AI Autofill".
|
||||
- This is highly effective for automating tags, description summaries, camera settings, and video transcripts.
|
||||
|
||||
### 3. Embedding Agent
|
||||
|
||||
A specialized utility agent that processes assets to generate search indexes.
|
||||
A specialized utility agent that processes assets to generate search indexes. You can create **only one** Embedding Agent.
|
||||
|
||||
- It runs embedding models on asset text, captions, or image descriptions.
|
||||
- Once created and enabled, Shumai automatically runs embedding generation for new file uploads if the file type is supported (image or video).
|
||||
- These search indexes are saved to power Shumai's natural language semantic search capabilities.
|
||||
- **Note:** The Embedding Agent is hardcoded to use `google-embedding-2` as the LLM model. Because of this, you cannot choose a provider or model. Make sure you set the `GEMINI_API_KEY` environment variable to make it work.
|
||||
|
||||
---
|
||||
|
||||
## Configuring an Agent
|
||||
|
||||
To edit or create an agent, navigate to **Team Settings** > **Agents**:
|
||||
To edit or create an agent, navigate to **Team Settings** > **Agents** (Note: Only Team Owners can change these settings):
|
||||
|
||||
- **Avatar**: Choose an avatar for the agent by picking from the predefined avatars.
|
||||
- **Enabled / Disabled Toggle**: Stop or start the agent. If the embedding agent is disabled, semantic search indexing is paused.
|
||||
- **System Prompt / Soul**: The core instructions that define the agent's behavior, guidelines, formatting requirements, and operational limits.
|
||||
- **Model Provider**: Select which registered provider and model the agent should use for execution.
|
||||
- **Model Provider**: Select which registered provider and model the agent should use for execution. _(Note: This option is disabled for Embedding Agents)._
|
||||
- **Assigned Skills**: Select which custom skills this agent is permitted to execute.
|
||||
|
||||

|
||||
_Placeholder: Configuration screen showing an agent's soul text editor, model dropdown, and assigned skills checkbox list._
|
||||

|
||||
|
||||
@@ -15,8 +15,7 @@ You can upload files and folders into Shumai using three methods:
|
||||
2. **Upload Buttons**: Click the **Upload File** or **Upload Folder** buttons in the toolbar.
|
||||
3. **Copy and Paste**: Copy an image or file from another website, application, or your clipboard, and paste it directly (`Ctrl+V` or `Cmd+V`) into the Shumai browser window to upload it instantly.
|
||||
|
||||

|
||||
_Placeholder: Dashboard interface displaying the drag-and-drop zone and active progress bar._
|
||||

|
||||
|
||||
---
|
||||
|
||||
|
||||
@@ -52,5 +52,4 @@ Define a custom sorting rule specific to the collection (e.g., sort by the custo
|
||||
- **Daily Deliverables**: A collection filtering files uploaded in the last 24 hours.
|
||||
- **Approved Assets**: A collection filtering assets matching `Tag` contains `marketing` AND `Approved` is `Yes`.
|
||||
|
||||

|
||||
_Placeholder: Interface modal of the Collection builder, displaying how filters are added and previewed._
|
||||

|
||||
|
||||
@@ -16,8 +16,7 @@ When reviewing video or audio files, comments can be anchored to specific second
|
||||
3. Click **Submit**. Your comment will be saved with a timestamp marker shown on the video timeline.
|
||||
4. Collaborators can click your comment card in the sidebar to jump the video player to the exact moment you commented on.
|
||||
|
||||

|
||||
_Placeholder: Screenshot of the video review interface showing comments linked to markers on the timeline._
|
||||

|
||||
|
||||
---
|
||||
|
||||
@@ -31,8 +30,7 @@ To give clear visual feedback, you can draw directly onto image or video frames:
|
||||
4. Type your comment and click **Submit**. Your drawing is saved along with your comment.
|
||||
5. When other users select your comment, your drawings will overlay on the player screen automatically at that frame.
|
||||
|
||||

|
||||
_Placeholder: Screenshot showing the drawing canvas overlay over a video frame with visual arrows and circles._
|
||||

|
||||
|
||||
---
|
||||
|
||||
|
||||
@@ -30,6 +30,8 @@ When building or editing your metadata schema, you can choose from these custom
|
||||
- **Checkbox (Toggle)**: A simple Yes/No switch for binary indicators (e.g., "VFX Required").
|
||||
- **Date**: A calendar date selector with support for custom display formats (Local, ISO, USA, Euro, Friendly) and optional time tracking.
|
||||
|
||||

|
||||
|
||||
---
|
||||
|
||||
## Default Built-in Fields
|
||||
|
||||
@@ -25,5 +25,4 @@ To use semantic search:
|
||||
2. Once active, Shumai will automatically begin indexing new uploads.
|
||||
3. Toggle the **Semantic** button next to the search input in the file explorer to search by conceptual meaning.
|
||||
|
||||

|
||||
_Placeholder: Screenshot of the search bar, emphasizing the active "Semantic" toggle option next to the text input._
|
||||

|
||||
|
||||
@@ -9,13 +9,15 @@ Shumai allows you to collaborate with external stakeholders who do not have acco
|
||||
|
||||
## Creating a Share Link
|
||||
|
||||
1. Hover over a file or folder and click the **Share** icon, or open the asset and click **Share** in the sidebar.
|
||||
2. Click **Create New Share Link**.
|
||||
3. Configure the settings described below and click **Generate**.
|
||||
4. Copy the URL and send it to your client.
|
||||
1. Select the files or folders you want to share, then right-click and choose **Create New Share Link**.
|
||||
2. Click the newly created **Share Link** in the project sidebar to open its settings page.
|
||||
3. Configure the share link in the right sidebar. You can:
|
||||
- Add password protection
|
||||
- Set an expiration date
|
||||
- Configure which metadata fields are visible to recipients
|
||||
4. Click **Copy Link** at the bottom of the sidebar, then send the link to your client or stakeholder.
|
||||
|
||||

|
||||
_Placeholder: Share modal settings including password input, expiration calendar, and field toggles._
|
||||

|
||||
|
||||
---
|
||||
|
||||
@@ -23,10 +25,9 @@ _Placeholder: Share modal settings including password input, expiration calendar
|
||||
|
||||
Once a share link is created, you can easily add more files or folders to it without generating a new link:
|
||||
|
||||
1. Locate the existing share link under the **Shares** section in the project sidebar.
|
||||
2. Select the files or folders you want to add from your file browser.
|
||||
3. Drag and drop them directly onto the share link item in the sidebar.
|
||||
4. The share link will immediately update to include the newly added assets.
|
||||
1. Select the files or folders you want to add from your file browser.
|
||||
2. Drag and drop them directly onto the share link item in the project sidebar.
|
||||
3. The share link will immediately update to include the newly added assets.
|
||||
|
||||
---
|
||||
|
||||
|
||||
@@ -11,8 +11,7 @@ As your projects grow, finding the right media files quickly is critical. Shumai
|
||||
|
||||
Searching is performed inside a dedicated **Search Dialog** that you can open by clicking the search icon in the toolbar.
|
||||
|
||||

|
||||
_Placeholder: Screenshot of the Search Dialog showing a keyword input, custom metadata criteria inputs, and the "Save as Collection" button._
|
||||

|
||||
|
||||
### Search Dialog Features
|
||||
|
||||
|
||||
@@ -55,5 +55,4 @@ Member management is handled using a unified **Members Dialog** popup:
|
||||
- **Add Team Members**: The dialog displays a list of active team members who are not yet added to this project. Select their project role (which will override their team role inside this project) and click **Add** to add them instantly.
|
||||
- **Invite External Collaborators**: Select a role (Editor or Reviewer) and click **Generate Link** to invite new users directly into this specific project.
|
||||
|
||||

|
||||
_Placeholder: Screenshot of the Members Dialog showcasing the list of active project members, the "Add Team Member" selection list, and the invite link generator._
|
||||

|
||||
|
||||
@@ -5,30 +5,32 @@ description: Configure video transcode strategies and preview proxy generation.
|
||||
|
||||
High-resolution camera formats and raw media files cannot be rendered natively in web browsers. Shumai transcodes these assets into web-friendly streaming formats in the background.
|
||||
|
||||
Regardless of the selected transcoding strategy, Shumai's pipeline will **always** generate at least one lightweight `180p` proxy version of the raw file to ensure fast scrubbing, thumbnail generation, and instant previews.
|
||||
|
||||
---
|
||||
|
||||
## Video Transcode Strategies
|
||||
|
||||
Team Owners can select how the transcoding engine handles uploaded files:
|
||||
Team Owners can select how the transcoding engine handles uploaded files. Shumai evaluates the original file's dimensions against its standard target resolutions (2160p, 1080p, 720p, 540p, and 360p) to determine the output.
|
||||
|
||||
### 1. Best Match (`best_match`)
|
||||
### 1. Best Match
|
||||
|
||||
- **Behavior**: Shumai checks the parameters of the uploaded file.
|
||||
- **Skip Transcoding**: If the file is already in a web-compatible format that browser engines can play directly (like standard MP4s), transcoding is skipped.
|
||||
- **Transcode**: If the file format is incompatible, it is transcoded. This is the recommended option as it saves processing time and storage space.
|
||||
- **Behavior**: Shumai checks the longest dimension (width or height) of the uploaded file. It then finds the highest standard resolution that is equal to or smaller than the original file, preventing unnecessary upscaling. If the original video is exceptionally small, it defaults to a minimum target of `360p`.
|
||||
- **Output**: Generates exactly two web-optimized files: the universal `180p` proxy and the single calculated "best match" resolution.
|
||||
- **Use Case**: This is the recommended option. It ensures web compatibility without wasting processing time and storage space on redundant lower-resolution versions.
|
||||
|
||||
### 2. Transcode All (`all`)
|
||||
### 2. Transcode All
|
||||
|
||||
- **Behavior**: Every single uploaded video is sent through the transcoding pipeline.
|
||||
- **Output**: Converts the file into a standardized, web-optimized MP4 profile. This ensures consistent streaming performance and compatibility across all client devices and browsers, regardless of the source.
|
||||
- **Behavior**: Shumai calculates the "best match" resolution exactly as it does in the previous strategy. However, instead of stopping at one version, it processes the video into the best match _and_ every standard resolution smaller than it.
|
||||
- **Output**: Generates a full cascade of files. For example, if the best match for an upload is `1080p`, Shumai will generate the `180p` proxy, `1080p`, `720p`, `540p`, and `360p` versions.
|
||||
- **Use Case**: Ideal for environments requiring adaptive bitrate streaming, ensuring consistent playback performance across all client devices regardless of varying network conditions.
|
||||
|
||||

|
||||
_Placeholder: Web UI Settings screen displaying the Media Processing tab with the strategy toggle selectors._
|
||||

|
||||
|
||||
---
|
||||
|
||||
## Processing Architecture
|
||||
|
||||
Transcoding is resource-intensive. Depending on your scaling needs, your administrator can run the transcode engine on the main server (Local Single-Server mode) or delegate tasks to specialized transcoding worker machines (Distributed High-Scale mode).
|
||||
Transcoding is resource-intensive. Depending on your scaling needs, your administrator can run the transcode engine on the main server (Single-Server mode) or delegate tasks to specialized transcoding worker machines (Temporal Worker mode).
|
||||
|
||||
_Note: In all modes, the processing machine must have the standard system media utilities (`ffmpeg` and `ffprobe`) installed._
|
||||
Note: In all modes, the processing machine must have the standard system media utilities (`ffmpeg` and `ffprobe`) installed.
|
||||
|
||||
@@ -11,8 +11,7 @@ When collaborating on creative assets, files naturally go through multiple itera
|
||||
|
||||
A **Version Stack** groups multiple revisions of the same file into a single entry. Instead of cluttering your folder with several similar files, Shumai collapses them into one item with a version badge (e.g., "V3") and a dropdown list of older versions.
|
||||
|
||||

|
||||
_Placeholder: File list view showing an asset with a version badge and a dropdown list of older versions._
|
||||

|
||||
|
||||
---
|
||||
|
||||
|
||||
|
After Width: | Height: | Size: 83 KiB |