LobeChat review

LobeChat, now presented as LobeHub, is a hosted team-agent workspace rather than just a chat interface for one model. It sits between workplace messaging, agent execution and review: teams connect chat apps, let agents pick up suitable requests, and check a delivery before accepting it. The homepage names Slack, Discord and Telegram as connected IM apps, and shows a GitHub pull-request workflow.

The product is aimed at moving work from conversation into tasks. Agents can run several tasks at once, pick up requests without being tagged, retain lessons from corrections and report completed work or decisions in a daily brief. The example workflow has an agent prepare a change, provide evidence for acceptance checks and merge a PR after approval. That makes LobeHub an orchestration layer for chat-driven agent work, not a replacement for your source host, model provider or team chat. Confirm which systems and permissions your intended workflow requires before adopting it.

The homepage describes integrations at the product level: connect Slack, Discord or Telegram, add agents to team chats, and use a GitHub workflow for changes and PRs. It also says users can plug in an IM app, agent or model. The showcased workflow includes acceptance checks, evidence, correction rounds and an approval before a PR merges.

The homepage does not document a developer API, SDK, supported programming language, CLI, webhook format or plugin interface. Those are key details for teams deciding whether to treat LobeHub as a component in their own stack or use it as a standalone workspace. Confirm API access, SDK support, authentication, event delivery, model-provider configuration and GitHub permissions in the vendor’s docs before planning an integration. The product’s chat connections and agent workflow are the named integration surfaces; they are not a substitute for a documented API contract.

The Free plan costs $0 and includes 500,000 credits a month, 15 MB of file storage, 100 vector entries, three active Artifact sites and one concurrent agent task. Each document can contain up to 100,000 characters. That is enough to evaluate a chat-agent workflow, but the single-task ceiling limits parallel use.

Plus costs $20 month to month or $18 per month with annual billing. It includes a weekly usage allowance, a five-hour usage limit, two concurrent tasks, 2 GB of file and cloud sandbox storage, and 10,000 vector entries. Pro is $100 monthly or $90 annual-equivalent, with five times Plus usage and five times its five-hour limit; Max is $200 monthly or $180 annual-equivalent, with ten times Plus usage and twenty times its five-hour limit. Their concurrency limits are five and ten tasks, respectively. LobeHub Cloud measures model use in Credits, and additional usage can be purchased. Ask the vendor how credits map to your models and expected workloads. Enterprise has custom pricing and lists private deployment options.

A Reddit post describes the earlier LobeChat as an open-source AI chat UI, but the current LobeHub homepage does not state the licence for this team-agent product. Confirm the current repository, licence terms and what code is available before relying on open-source rights or planning a fork.

The Enterprise Edition lists custom pricing, a commercial licence, self-hosted Provider and private models, as well as private deployment. That is a paid route to deployment control, not evidence that every hosted-plan feature can be self-hosted. The homepage does not describe data export or a migration path. Before storing operational memory, task history or artifacts, ask how to export them and whether they can be used outside LobeHub. Chat integrations and GitHub work may reduce workflow lock-in, but agent memory, credits and product-specific task records could still require migration planning.

Capacity is plan-bounded: concurrency ranges from one task on Free to ten on Max, while storage, vector entries, Artifact sites and document size also have tier limits. Paid plans have weekly allowances and five-hour usage limits; the higher tiers multiply those allowances. The pricing page does not publish API rate limits, an uptime SLA or incident terms. Ask about those, along with data retention, access controls, audit logs and compliance, before putting sensitive or business-critical work through agents.

The homepage presents approval controls for PR merges and says deployments need approval in its example, but teams should verify how those permissions work in their configuration. Reddit reports about the earlier LobeChat chat UI are mixed: one LocalLLaMA poster said a Docker Compose deployment was quick and praised web search, while other users discussed competing tools and governance preferences. Those comments concern the chat product, not evidence about the current team-agent service’s reliability or security.