Some links on this page are affiliate links: if you buy through them we may earn a commission, at no extra cost to you.
Assist is Home Assistant’s user-facing voice and text pipeline, a script is a saved sequence of actions that runs only when something calls it, and an LLM conversation agent is a language-model-based interpreter that can choose tools, including eligible scripts. You do not need an LLM to use Assist: its built-in conversation agent can match supported sentences locally. The three components can work together, but they solve different problems. Understanding the boundaries helps you choose the right mechanism, expose only the controls you intend, and troubleshoot the correct layer when a request fails.
What is the difference between Assist, a script, and an LLM agent?
Think of the three as different layers. Assist carries a person’s request through a processing pipeline. A conversation agent interprets the resulting text. A script performs a predefined sequence after a caller starts it. An LLM conversation agent is one kind of conversation agent; it uses a language model to interpret requests and may call tools that Home Assistant exposes.
As an Amazon Associate I earn from qualifying purchases.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREEClear out junk files and repair common Windows errorsFree Scan →| Component | Primary role | How it starts | LLM required? | Typical result |
|---|---|---|---|---|
| Assist | User-facing voice or typed assistant and processing pipeline | A voice or typed request enters an Assist pipeline | No. The built-in agent can match sentences locally | Text or speech is interpreted and an intent or tool call is produced |
| Script | Reusable, saved sequence of actions | A dashboard, Assist, automation, or another caller explicitly runs it | No inherent LLM requirement | Actions run in the order and logic defined by the script |
| LLM conversation agent | Flexible conversational interpretation and tool selection | A conversation request is sent to the configured agent | Yes, by definition | The model responds and may call exposed entities or tools |
A useful rule is: Assist is the route, the conversation agent is the interpreter, and the script is the reusable work. Calling a script through Assist does not turn the script into an LLM, and installing an LLM integration does not replace the Assist pipeline.
How does an Assist request travel through Home Assistant?
Assist can accept typed text or a spoken request. A voice pipeline may include wake-word detection, speech-to-text, conversation processing, and text-to-speech. Depending on your configuration, individual stages can run on the device, through an integration, or in the cloud.
#1 Best Overall
- Read Before You Buy — No Video Output: These adapters support charging and USB 2.0 data transfer, but cannot transmit video signals. Except for standard USB webcams (which use USB data only), they are not compatible with HDMI/DisplayPort cables, video-capable USB-C hubs, or docking stations with video output.
- Convert USB-A Ports to USB-C: Designed to connect USB-C earphones, cables, flash drives, card readers, and other USB-C accessories to standard USB-A ports. Plug-and-play with no drivers or software required.
- Aluminum Alloy Housing: Built with a sturdy aluminum alloy shell that aids in heat dissipation and protects against daily wear and scratches. Designed to maintain a stable and secure connection.
- Compact & Travel-Friendly: The ultra-compact design allows the adapter to stay plugged into your device without blocking adjacent ports or adding bulk, reducing wear and tear on your original USB ports.
- 12-Month Warranty: Backed by a 12-month manufacturer warranty for peace of mind. Designed to meet strict quality control standards for reliable everyday performance.
- Input: You type a request or speak after a wake word or Assist button.
- Speech recognition (voice only): Audio is converted to text by the selected speech-to-text provider.
- Conversation processing: The selected conversation agent interprets the text. The built-in agent uses known sentence patterns and intents; an LLM agent uses a model and its available tools.
- Home control: The agent may call an exposed entity, service, or script. A script then executes its own action sequence.
- Response: Home Assistant returns text and, in a voice pipeline, can pass it to text-to-speech.
Assist was introduced in Home Assistant 2023.2. The introduction date identifies the feature’s release context, not a promise that every pipeline stage or integration behaves identically across versions.
Does Assist require an LLM?
No. The built-in conversation agent can handle supported commands with local sentence matching, so lights, covers, media players, and other supported intents can work without sending the request to a language model. An LLM is an optional alternative when you want broader conversational interpretation.
When the built-in agent is a good fit
- Commands follow recognizable Home Assistant sentence patterns.
- You prefer predictable intent matching and local processing.
- You want to minimize model context, external calls, or tool exposure.
When an LLM agent helps
- A request is phrased in a less predictable, conversational way.
- The assistant must decide among several tools based on context.
- You want a model to handle follow-up language or explain what it did.
LLM behavior is model-dependent. A more flexible interpreter can also be less deterministic, so expose only the entities and tools needed for the task and give each callable script a precise description.
What exactly is a Home Assistant script?
A script is a saved sequence of actions. It has no trigger of its own: it runs only when something starts it. A dashboard button, an automation, Assist, or another script can be that caller.
Use a script for reusable or on-demand work
- Run a “movie time” sequence that dims lights, closes blinds, and starts a player.
- Provide one routine that several dashboards and automations can call.
- Start a multi-step operation manually when you decide the timing.
Use an automation when a state change should start the work
If the sequence should begin when a motion sensor changes, a schedule fires, or a person arrives, an automation supplies the trigger and can call a script for the reusable actions. Separating the trigger from the action sequence keeps the script callable from other places.
Rank #2
- 5-in-1 USB-C Hub: Experience comprehensive connectivity featuring a Power Delivery input, two USB-A 2.0 ports, a USB-A 3.0 port, and an HDMI port. (Note: The USB-C power delivery input port is only for connecting an external wall charger to power your laptop and cannot power peripheral devices.)
- 90W Pass-Through Charging: Achieve optimal charging with 90W pass-through power to your laptop, supported by a total input of 100W, with the hub reserving 10W for operational efficiency. (Note: Wall charger not included.)
- Quick Data Transfers: Accelerate your productivity with rapid data transfers using a high-speed 5Gbps USB 3.0 port and two 480Mbps USB 2.0 ports.
- 4K HDMI Display: Enhance your visual experience with a hub capable of delivering 4K resolution at 30Hz in both mirror and extend modes. Please note that this hub is compatible with MacBook (macOS 12 and newer), Windows 10 and 11, ChromeOS, and laptops equipped with DP Alt Mode and Power Delivery. Note: This device is not compatible with Linux.
- What You Get: Anker USB-C Hub (5-in-1, 4K HDMI), welcome guide, 18-month warranty, and our friendly customer service.
Use a scene for a saved group of states
A scene applies a set of device states together. It is usually a better fit for “set these lights and shades to these values” than for delays, branching, loops, or a long sequence of steps.
Do these 3 things before closing this tab:
1Fix the driver behind crashes, sound loss and screen glitches2Clear out junk files and repair common Windows errors3Scan for outdated or missing drivers - takes under a minuteHow does an LLM conversation agent use Home Assistant?
Home Assistant’s LLM framework supplies tools to an API, while a conversation integration configures an agent to use a particular model. The agent receives conversational text and can select tools that Home Assistant and the integration make available.
Scripts exposed to an LLM are not shown as ordinary static entities. Home Assistant describes them as callable tools. The model needs a concise description that says what the script does and when it should be called. If the script accepts inputs, describe each input’s meaning and expected format so the model can supply usable values.
A script can also stop with a success or error message. That result gives the agent something explicit to relay instead of forcing it to infer whether the operation worked.
The LLM integration documentation was introduced in Home Assistant 2026.7. That date refers to the framework integration page; it should not be read as the first possible date for every third-party or built-in LLM conversation integration.
Outdated Drivers Are Slowing You Down
One free scan finds every outdated or missing driver and matches the right update for your exact hardware.Free scan · exact hardware matchWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallRank #3
- Sleek 7-in-1 USB-C Hub: Features an HDMI port, two USB-A 3.0 ports, and a USB-C data port, each providing 5Gbps transfer speeds. It also includes a USB-C PD input port for charging up to 100W and dual SD and TF card slots, all in a compact design.
- Flawless 4K@60Hz Video with HDMI: Delivers exceptional clarity and smoothness with its 4K@60Hz HDMI port, making it ideal for high-definition presentations and entertainment. (Note: Only the HDMI port supports video projection; the USB-C port is for data transfer only.)
- Double Up on Efficiency: The two USB-A 3.0 ports and a USB-C port support a fast 5Gbps data rate, significantly boosting your transfer speeds and improving productivity.
- Fast and Reliable 85W Charging: Offers high-capacity, speedy charging for laptops up to 85W, so you spend less time tethered to an outlet and more time being productive.
- What You Get: Anker USB-C Hub (7-in-1), welcome guide, 18-month warranty, and our friendly customer service.
How do I let an LLM call a script?
Expose only the scripts that the agent genuinely needs. The setup is opt-in, and Home Assistant documents a maximum of 128 scripts and tools in total.
- Configure a supported conversation integration. Select its Assist mode and enable home-control capability as required by that integration.
- Open the script’s settings. Enable the script for Assist so it can be registered as a callable tool.
- Write a specific description. State the outcome and the circumstances for calling it, such as “Use this to prepare the living room for a movie; do not use it for ordinary light changes.”
- Document inputs. Explain units, allowed values, and whether an input is optional. For example, say whether a duration is seconds or minutes.
- Limit exposure. Disable unrelated scripts and entities. A smaller tool set gives the model less context to search and reduces accidental matches.
- Test the tool call. Use the Assist Debug view for the LLM pipeline and verify that the expected tool was selected, arguments were valid, and the script returned the intended result.
Exposing a script does not make it a normal entity in the model’s static entity list. It registers a callable tool, so a failure may appear as a tool-selection or argument problem rather than an entity-state problem.
What is the role of entity exposure?
Entity exposure controls what an assistant can see and operate. It is opt-in, so you can limit an Assist or LLM agent to the lights, climate devices, locks, scripts, and other controls relevant to its job.
- Safety: Do not expose a lock, garage door, alarm, or other sensitive control unless the use case requires it and you understand the integration’s confirmation behavior.
- Clarity: Distinct names and descriptions reduce ambiguity when several devices have similar labels.
- Context size: More exposed entities and tools increase the matching work for the built-in agent and the context supplied to an LLM.
- Maintenance: Review exposure after adding rooms, devices, or scripts; an old permission can remain active after the original use case changes.
How do Assist, scripts, and LLM agents work together?
A typical voice request looks like this: Assist receives “prepare the living room,” the selected conversation agent interprets it, and the agent calls an exposed script that dims lights, adjusts shades, and starts a device. The script remains a deterministic sequence; the agent decides whether it is the right tool.
The reverse direction is also possible. An automation or script can use the conversation.process action to send text to a selected or default conversation agent and capture its response. This is useful when another part of your setup produces text that should be interpreted as a request, rather than when you simply need to run fixed actions.
Rank #4
- Dual Converters, Infinite Potential：Includes 2× USB C male to USB A female adapters and 2× USB A male to USB C female adapters. Perfect for a wide range of uses—tablets with Bluetooth keyboards, expand USB ports on macbook, and more. Two different converters for all your daily needs
- Next-Level 10Gbps & 3A Charging: No more slow 480Mbps, this usb to usb c adapter has a transfer speed of up to 10Gbps, allowing you to do more transferring in less time. This usb adapter fits both USB A and USB C charger, supporting up to 3A fast charging
- Upgraded Exquisite Craftsmanship: With an aluminum alloy housing and metal connector, the usbc to usb adapter is extremely durable and sturdy. Rigorously tested to withstand more than 10,000 times of plugging and unplugging, ensuring long-lasting performance
- Broad Compatible: The usb c to usb adapter widely supports all USB C/ USB A devices like laptops, tablets, cellphones, car chargers, and phone chargers. Such as compatible with MacBook Pro/Air 2023/2022, Thunderbolt 4/3 Devices,Apple MagSafe Watch 9/8/7/SE/Ultra, iPad Pro 2022/2021, Samsung Galaxy S23/S20/S10, and iPhone 17/16/15 Pro. Plug and play
- Please Note: To reach 10Gbps speed, keep the cable under 3.3 ft. For USB A Male to USB C adapters, try flipping the USB C connector. USB C Male to USB A adapters support bidirectional 10Gbps transfer within 3.3 ft
Voice hardware is optional. A phone, browser, or other client can provide the interface. For a dedicated satellite, Home Assistant’s architecture uses microphone and speaker hardware while Home Assistant runs the pipeline; ESP32-based satellite tutorials and the Home Assistant Voice Preview Edition are examples of that approach. Check current compatibility before buying a board because hardware support and listings can change.
Which mechanism should I choose?
| If your requirement is… | Choose… | Reason |
|---|---|---|
| A person should speak or type a request | Assist | It provides the user-facing pipeline and routes the request to a conversation agent |
| Predictable supported commands with local matching | Built-in conversation agent | No LLM is required and intent matching is constrained |
| Flexible wording or conversational tool selection | LLM conversation agent | The model interprets text and can choose exposed tools |
| A reusable sequence started on demand | Script | It packages actions without requiring its own trigger |
| A sequence started by a sensor, schedule, or state change | Automation, often calling a script | The automation supplies the trigger while the script supplies reusable actions |
| A simultaneous set of target states | Scene | Scenes are designed to apply saved states rather than run procedural logic |
Many installations use more than one row: Assist is the entry point, the built-in or LLM agent interprets the request, and scripts or scenes perform the selected work.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How should I test a request when something fails?
Test the layer that is actually failing; the developer tools do not all exercise the same behavior.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Testing built-in sentence matching
Open Settings > Tools > Assist. This developer Assist tab tests the built-in sentence matcher. It does not execute actions, so a successful match there does not prove that a script, service call, or device will run.
Testing an LLM pipeline
Use the Assist Debug view for the pipeline that uses the other conversation agent. Inspect the request, selected tool, arguments, returned result, and final response. Because scripts are registered as tools rather than static entities, look for an action or tool call instead of searching only for an entity state.
Best Value
- 5-in-1 Connectivity: Equipped with a 4K HDMI port, a 5 Gbps USB-C data port, two 5 Gbps USB-A ports, and a USB C 100W PD-IN port. Note: The USB C 100W PD-IN port supports only charging and does not support data transfer devices such as headphones or speakers.
- Powerful Pass-Through Charging: Supports up to 85W pass-through charging so you can power up your laptop while you use the hub. Note: Pass-through charging requires a charger (not included). Note: To achieve full power for iPad, we recommend using a 45W wall charger.
- Transfer Files in Seconds: Move files to and from your laptop at speeds of up to 5 Gbps via the USB-C and USB-A data ports. Note: The USB C 5Gbps Data port does not support video output.
- HD Display: Connect to the HDMI port to stream or mirror content to an external monitor in resolutions of up to 4K@30Hz. Note: The USB-C ports do not support video output.
- What You Get: Anker 332 USB-C Hub (5-in-1), welcome guide, our worry-free 18-month warranty, and friendly customer service.
Common failure branches
- The sentence matches but nothing happens: You tested the built-in matcher, which intentionally does not execute actions. Test the actual pipeline or call the script directly.
- The LLM never selects the script: Confirm that the script is enabled for Assist, the integration has home-control mode enabled, and the description clearly distinguishes it from similar tools.
- The model calls the wrong routine: Reduce exposure, rename ambiguous tools, and describe when each script should not be used.
- Arguments are invalid: Add units, allowed formats, and examples to the input-field descriptions; then inspect the debug trace for the values the model sent.
- A hard-limit error appears: Count exposed scripts and tools. The documented total limit is 128, so disable tools that are not needed.
What are the practical trade-offs?
Predictability versus flexibility
The built-in agent is constrained by known intents, which makes routine commands easier to reason about. An LLM can understand more varied language but introduces model-dependent decisions and requires careful tool descriptions.
Local processing versus external services
Assist pipelines can combine local and cloud stages. Choosing local speech or conversation processing can reduce external data transfer, while cloud or model-backed stages may offer broader language support. The exact boundary depends on the pipeline and integrations you select.
Small tool sets versus broad control
Exposing everything may appear convenient, but it increases ambiguity and context. A narrowly scoped agent with clearly named entities and scripts is easier to audit and safer to operate.
What should I remember?
Start with the user action you need. Choose Assist for the interaction, a built-in agent for constrained local intent matching, an LLM agent for flexible interpretation, a script for reusable procedural actions, an automation for trigger-driven behavior, and a scene for a saved group of states. Then expose only the controls that support that purpose and verify the real pipeline in Assist Debug rather than assuming a sentence match executed anything.
Frequently Asked Questions
Can Home Assistant Assist work without internet access?
Some pipeline stages can run locally, including the built-in conversation matching and locally configured speech components. Offline operation is not automatic for every setup: each wake-word, speech-to-text, conversation, and text-to-speech stage must use a local option if you want to avoid cloud services.
Can an LLM control a lock or garage door through Assist?
Only if that entity or a script controlling it is exposed to the agent. Treat sensitive controls separately, expose them only when necessary, and verify the integration’s confirmation and authentication behavior before allowing conversational control.
Why does my script appear in Assist but not in the LLM’s entity list?
Scripts are registered as callable tools, not ordinary static entities. Check the LLM integration’s tool or action trace and make sure the script is enabled for Assist; searching only the entity list will miss it.
Do I need dedicated voice hardware to use Assist?
No. A phone, browser, or another client can provide typed or spoken input. Dedicated satellites add microphones and speakers for hands-free use, but they are an optional interface rather than a requirement for the Assist software pipeline.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
