What Is LightVela

LightVela is a cloud-hosted AI Agent platform built by the Tencent Cloud team. It takes the open-source Hermes Agent and runs its deployment, operations, and 24/7 around-the-clock execution entirely in the cloud. You don't need to provision a server or learn any DevOps — just click through a web interface to spin up your own personal AI assistant, then connect it to WeChat, QQ, WeCom (Enterprise WeChat), Feishu, and DingTalk conversations. LightVela supports multiple models, skill extensions, persona injection, scheduled tasks, and cross-session memory, and it can host several Agents at the same time.

Key Features

• Cloud hosting — Deployment, operations, and 24/7 execution all happen in the cloud. No server or DevOps knowledge required; configure everything from a web page.
• Multi-channel access — Connect five major IM apps: WeChat, QQ, WeCom (Enterprise WeChat), Feishu, and DingTalk. Scan a QR code to authorize, then chat with your Agent right inside the messaging app.
• Multiple models — Plug in mainstream domestic models such as DeepSeek, Kimi, Doubao, MiniMax, and Qianwen. Three integration methods are supported: a monthly Coding Plan, pay-as-you-go API, and any OpenAI-compatible endpoint.
• Skill extensions — Install individual skills or skill packs from Tencent's SkillHub, and the Agent decides when to invoke them based on context.
• Persona injection — Give your Agent a distinct personality and voice through three methods: random persona box, persona testing, or uploading a persona file.
• Scheduled tasks — Run AI smart tasks plus notification and script tasks that don't consume credits. Chain tasks together by linking an upstream task.
• Memory accumulation — Automatically build up Agent memory and a user profile that persists across sessions, so it gets to know you better the more you use it.
• Multi-Agent hosting — Create several Agents, each in its own isolated cloud environment, with separate channels, models, skills, personas, and memory.

How to Use LightVela

• Sign up and log in — Visit the official site at https://lightvela.com/ to register and sign in.
• Create an Agent — Pick an avatar, name it, add a description, and create your own Hermes Agent with one click.
• Configure a model — It ships with trial credits out of the box; you can also connect your own API key, a Coding Plan, or an OpenAI-compatible endpoint.
• Connect a channel — In channel settings, scan to authorize any of WeChat, QQ, WeCom, Feishu, or DingTalk, or just use the web interface directly.
• Install skills — Search SkillHub and tick the skills or skill packs you need; the Agent calls them automatically based on the conversation.
• Inject a persona — Use the random box, persona testing, or upload a persona file to set the Agent's character and style.
• Set scheduled tasks — Choose a template for AI smart, notification, or script tasks, and link an upstream task to form a chain.
• Start chatting — Send tasks to your Agent in the IM app or on the web; the cloud runs them and returns results automatically.

LightVela's Core Advantages

• Always-on cloud presence — Your Agent runs 24/7, so tasks execute without your local device being on, freeing you from desktop-client dependency.
• Native Tencent IM integration — Scan to connect WeChat, QQ, WeCom, Feishu, and DingTalk; this entry-point advantage is hard for other products to copy.
• Long-term cross-session memory — Automatically accumulates MEMORY.md and a user profile, so the Agent knows you better over time instead of re-introducing itself every session.
• Independent multi-Agent hosting — Raise multiple "digital employees" with isolated resources and independent configurations for teamwork and division of labor.
• Free model choice — Compatible with mainstream domestic models and any OpenAI endpoint, so you're not locked into a single vendor.
• SkillHub ecosystem — Backed by Tencent's skill library for one-click skill-pack installs, with virtually unlimited capability expansion.

LightVela vs. Similar Products

Use Cases

• Personal knowledge management and topic research — Let the Agent continuously gather the latest content in your field, store it in a knowledge base, and summarize it to support content planning and decisions.
• Self-media / official account operations — Configure writing and layout skill packs; give a topic and the Agent drafts outlines, first drafts, or layout assets in the cloud, with scheduled tasks automating daily morning/evening reports.
• Coding partner and automated ops — For developers: configure PRD writing and code-assist skills, and use credit-free script tasks to run deterministic checks on a schedule.
• Finance / business assistant — After installing finance skills, it periodically pulls market and earnings data and generates analysis reports, running 24/7 in the cloud and pushing alerts to WeChat when triggers fire.
• Life manager — Set notification tasks like reminders to drink water, rest, or handle schedule items; combined with persona injection and long-term memory, the Agent adapts to your preferred communication style.

FAQ

Q: Does LightVela require me to provide my own server? A: No. LightVela hosts the Agent's deployment, operations, and 24/7 execution entirely in the cloud. You need no server and no DevOps knowledge — just click through the web interface to create and configure your own Agent.

Q: Which chat apps can LightVela connect to? A: It supports five major IM apps — WeChat, QQ, WeCom (Enterprise WeChat), Feishu, and DingTalk. After scanning to authorize, you can chat with the Agent directly inside the app; you can also use the web interface.

Q: Can I use my own model or API key? A: Yes. LightVela works with mainstream domestic models such as DeepSeek, Kimi, Doubao, MiniMax, and Qianwen, and supports three connection methods — a monthly Coding Plan, pay-as-you-go API, and any OpenAI-compatible endpoint — so you're not tied to a single vendor.