LobeHub
LobeHub is an open-source AI agent orchestrator that runs a 24/7 multi-agent team under one Chief Agent Operator.
If your bottleneck is coordination rather than raw model quality, LobeHub is one of the few open-source platforms actually built for that. The credit system means your real bill tracks model choice, so heavy Claude Fable 5.1 or GPT-6 Astra use can outrun the sticker price fast. Pick it when you want agents working overnight and reporting into Slack; skip it if one good chatbot answers your questions.
Verified 7d ago · liveness 87/100 · cite: rightaichoice.com/tools/lobehub
- Teams with long-horizon, repetitive workloads they want agents to run overnight
- Developers who want an open-source, self-hostable multi-agent orchestration platform
- Organizations moving several AI tools under one agent system with shared memory
- Teams already in Slack, Discord, or Telegram that want agent output delivered there
- Anyone who just wants a single-LLM chatbot without orchestration overhead
- Teams without DevOps support who can't run or tune a self-hosted deployment
- Budget planners who need a fixed monthly bill instead of credit-metered model spend
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
Skip LobeHub if you need a simple single-LLM chatbot without agent orchestration, if you lack technical support for setup and tuning, or if you require predictable flat pricing instead of credit-based metering.
Credit consumption varies wildly by model: GPT-5.5 Pro costs 30M credits per 1M input tokens vs. GLM-5.3-Flash at 0.075M, so a single high-end model call can eat a huge chunk of your monthly quota.
LobeHub's credit-based pricing is best for teams that can estimate their model usage and want a single platform for many AI models. At $9.9/mo Starter (or $19.9/mo Premium), it's cheaper than using each model's native API separately, but it may cost more than a single-model subscription like ChatGPT Plus ($20/mo) if you only need one model. For heavy multi-agent workloads, the Ultimate tier at $39.9/mo offers 35M credits, which is competitive with direct API costs but requires careful tracking.
In short
LobeHub — LobeHub is an open-source AI agent orchestrator that runs a 24/7 multi-agent team under one Chief Agent Operator. Best for Teams with long-horizon, repetitive workloads they want agents to run overnight, Developers who want an open-source, self-hostable multi-agent orchestration platform, Organizations moving several AI tools under one agent system with shared memory. Free to start; paid plans from $9.9/mo.
What's new in LobeHub
Checked 5 days agoAcross the latest 10 updates: 1 pricing change, 2 changelog entries and 7 news mentions.
Lobe Eval: testing agents on real-world tasks
LobeHub's Lobe Eval runs agents through the real product execution path in a resettable tool environment, scoring runs with traces and Pass³.
LOL #18: Agentic AI Summit at Berkeley, Pi joins the bench
LobeHub sponsored Agentic AI Summit 2026 at UC Berkeley; Pi is now connectable as a coding agent alongside Codex, Claude Code, Amp and OpenCode.
LobeHub v2.2.12: Home dashboard and local Agent setup
Home becomes an agent dashboard; desktop gains local MCP connector installs, a single Agent setup wizard, Markdown agent profiles and Kimi K3 reasoning effort controls.
LOL #17: ChatGPT as a provider, Kimi K3 Fast
Connect a ChatGPT subscription as a provider with no API key; Kimi K3 gets a low-latency Fast variant; three frontier models build the same game.
LobeHub v2.2.11: Gemini 3.6 Flash support
Gemini 3.6 Flash added with four thinking levels, multimodal input, web search and tool use; desktop terminal gains multi-session tabs.
LOL #16: Kimi K3, Grok 4.5 and a terminal in the Lab
Kimi K3 and Grok 4.5 land in the model picker, and LobeHub ships a built-in terminal as an opt-in Lab experiment.
How an OAuth phish hijacked LobeHub's X account
LobeHub details the OAuth phishing attack that took over its X account, how the attack worked and the response.
LobeHub v2.2.11: WeChat messaging moves to paid plans
WeChat message channel now requires a paid personal or Pro/Business Workspace plan, citing channel stability and operating costs.
LOL #15: Fable 5 returns, GLM-5.2 Fast, v2.2.9 updates
Anthropic's Fable 5 is available again in LobeHub, GLM-5.2 Fast arrives, and v2.2.9 adds platform quality-of-life updates.
LOL #14: AWS Summit China, new mobile look, GLM-5.2
LobeHub founders presented at AWS Summit China 2026; the mobile app was redesigned and GLM-5.2 became one of the platform's most used models.
Viability Score
How well maintained and how widely used is LobeHub? Built from what the vendor actually publishes (docs, changelog, tutorials, integrations, pricing), whether the site is live, and how much real users discuss it. How we calculate this
Last calculated: September 2026
How we score →Key Features
- Chief Agent Operator (CAO) hires, schedules, and reports on a 24/7 agent team
- Auto agent formation from a one-sentence description
- Agent Groups with auto team formation and parallel multi-task execution
- Skills Marketplace with 334,081+ skills
- MCP marketplace with 100,774+ MCP servers
- Multi-model access to Claude Fable 5.1, GPT-6 Astra, Grok 4.6, Kimi K3, GLM-5.3
- Gemini 3.6 Flash support with four thinking levels and web search
- IM Gateway deployment to Slack, Discord, Telegram, Feishu, WeChat, LINE, QQ
- Personal memory with continual learning and white-box editable memory
- Multimodal workflow across Pages, Schedule, Project, and shared Workspace
- RAG, file upload, and knowledge management
- Image and video generation, plus SVG and interactive HTML rendering
- Built-in desktop terminal with tabs and MCP connectors
- Home dashboard with agent status cards and onboarding starter tasks
- Self-hosting on Windows, macOS, Linux, or Docker
About LobeHub
LobeHub turns a pile of AI chats into an operating team. Its Chief Agent Operator (CAO) hires, schedules, and reports on specialized agents that run in parallel around the clock, so long-horizon work keeps moving while you're offline. The pitch is aimed at developers, ops leads, and small product teams who already juggle several models and channels and want one roof over all of it. Setup is deliberately thin. Describe a task in a sentence and LobeHub auto-configures agent names, roles, skills, and behaviors; the vendor page now lists 334,081+ skills and 100,774+ MCP servers in its marketplace. Agent Groups add auto team formation, parallel multi-task execution, and iterative improvement through feedback loops, while Pages, Schedule, Project, and Workspace carry multimodal work with shared context. Personal memory includes continual learning and white-box, editable memory you can inspect rather than trust blindly. Agents can also live where your team already talks. The IM Gateway deploys them into Slack, Discord, Telegram, Feishu, WeChat, LINE, and QQ, and recent releases added a home dashboard with agent status cards, a local agent setup wizard, MCP connectors in the desktop app, and terminal tabs. Self-hosting runs on Windows, macOS, Linux, or Docker, with Enterprise offering private deployment, brand theming, and user management. On price, LobeHub Cloud is credit-metered: Free gets 500,000 credits/month, Starter $9.9/month yearly, Premium $19.9, Ultimate $39.9, each unlocking bigger model access like GPT-6 Astra and Claude Fable 5.1. It is a heavier commitment than a hosted chatbot, and that is the honest trade.
Behind the Verdict
Most "multi-agent" tools are a chat window with a planner glued on. LobeHub's CAO is a different shape: you hand over a goal, it assembles the team, runs tasks in parallel, and posts results back to Slack or Discord while you sleep. The 500-issue sweep demo on the vendor page is not subtle marketing — it describes the actual workflow the product is built around. We'd reach for this when work is long-horizon and repetitive: triaging an issue backlog, monitoring a storefront, running scheduled research, or keeping several agents on separate projects with shared context. The Skills Marketplace and MCP server count matter here, because the ceiling on what an agent can do is really the ceiling on what it can connect to. Where it bites is cost predictability. Credits are consumed per token, and the model table shows the spread: DeepSeek V4.1 Flash outputs at 1.32M credits per 1M tokens versus 50M for Claude Fable 5.1 or GPT-6 Astra. A Free plan's 500,000 monthly credits buys roughly 600 GLM-5.3-Flash messages and nothing at all from the premium tier. Budget-conscious teams should test on cheap models first and only promote the agents that earn it. The closest alternative depends on what you want. Poe and ChatGPT are hosted, simple, and won't ask you to think about orchestration; that simplicity is real value if you have one recurring question a day. Frameworks like CrewAI or AutoGen give developers more low-level control but leave you building the scheduling, reporting, and channel delivery yourself. LobeHub's edge is that those parts already exist. Two practical cautions before you commit. First, WeChat connections require a paid personal plan or a Pro/Business Workspace for new connections after July 10, 2026 — if WeChat is your primary channel, factor that in. Second,
Researching LobeHub? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Real-world workflow fit
Concrete scenarios for the personas LobeHub actually fits — and what changes day-one when you adopt it.
You have 500 open issues and want to categorize and prioritize them overnight.
Outcome: You describe the task in one sentence, CAO dispatches 50 agents that work in parallel, and by morning you have a summarized report with suggested priorities, delivered to your Slack.
You want to answer common support questions 24/7 without staff monitoring.
Outcome: You set up an agent in Discord via the IM Gateway, attach a knowledge base, and it handles tickets, escalating complex ones to humans, all with delivery checks.
You need to compare output quality of multiple models for a writing task.
Outcome: You use LobeHub's multi-model support to run the same prompt across GPT-5.6, Claude Sonnet 5, and Gemini 3.6 Flash, viewing results side-by-side in one interface.
Use Cases
- Self-host a polished ChatGPT-style UI on Vercel and bring your own API keys.
- Compare GPT-5.6, Claude Sonnet 5, and Gemini 3.6 Flash side by side from one interface.
- Install community-built agent skills (translator, code reviewer) instead of writing system prompts.
- Run Ollama locally and attach LobeHub as the front-end for a fully-private setup.
- Use agent groups for stock trading analysis, meeting summaries, or paper reviews.
- Assign tasks to agents with dependencies and cron schedules using Agent Tasks.
- Let CAO autonomously sweep 500 GitHub issues overnight with 50 agents.
- Deploy agents in Slack or Discord to answer support tickets 24/7.
Models Under the Hood
as of 2026-09-15
Limitations
- LobeHub Cloud is credit-based: the Free plan provides 500,000 credits per month (~600 messages on GLM-5.3-Flash) with only 10 MB of file storage, and several models (GPT-6 Astra, Claude Fable 5.1, Kimi K3) are not supported on the Free tier.
- Higher message volumes, vector storage, and agent memory require paid plans ranging from $9.9 to $39.9 per month.
- After July 10, 2026, new and existing WeChat messaging connections require a paid personal plan or Pro/Business Workspace.
- Some capabilities such as the desktop terminal and local MCP connector installation depend on the desktop app.
as of 2026-08-30
Verification history
We have re-verified LobeHub 19 times since . Each pass re-reads the vendor's own pages and re-checks every listed field against that evidence; passes where nothing had changed are marked as such.
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
- — re-verified summary, description, our verdict, our analysis, pricing model, pricing tiers, features, integrations, who it suits, who should skip it
Showing the 6 most recent of 19 verification passes.
Free to cite with attribution — this page re-verifies continuously.
12-month cost
Project the real annual outlay, including the implied monthly cost when only an annual tier is published.
Vendor list price only. Add-on usage, seat overages, and contract minimums are surfaced under Hidden costs & gotchas.
Plans compared
For each published LobeHub tier: who it actually fits, and what it adds vs. the previous tier. Cross-reference the cost calculator above for projected annual outlay.
Free
$0/mo
Ideal for
Casual users who want to try out basic AI chat and image generation without paying, and are okay with limited models and storage.
What this tier adds
Starting tier: 500K credits/month, DeepSeek V4 Pro and Grok 4.6 only, 10 MB file storage, unlimited Pages.
Starter
$9.9/mo billed yearly ($12.9/mo monthly)
Ideal for
Individuals who need more than the free tier and want access to Claude and GPT models, with 5M credits/month for moderate use.
What this tier adds
Adds Claude Sonnet 5, Opus 5, and GPT-5.6 access, 1 GB file storage, 5,000 vector entries, Agent Memory.
Premium
$19.9/mo billed yearly ($24.9/mo monthly)
Ideal for
Power users who need 3x the credits, priority email support, and more storage, at a price that still fits individual budgets.
What this tier adds
Adds 15M credits/month, 2 GB file storage, 10,000 vector entries, and priority email support.
Ultimate
$39.9/mo billed yearly ($49.9/mo monthly)
Ideal for
Professional users or small teams with heavy AI usage who want the most credits and priority chat support.
What this tier adds
Adds 35M credits/month, 4 GB file storage, 20,000 vector entries, and priority chat plus email support.
Enterprise Edition
Custom (contact sales)
Where the pricing makes sense
The company stage and team size where LobeHub's pricing actually pencils out — and where peers do it cheaper.
LobeHub's credit-based pricing is best for teams that can estimate their model usage and want a single platform for many AI models. At $9.9/mo Starter (or $19.9/mo Premium), it's cheaper than using each model's native API separately, but it may cost more than a single-model subscription like ChatGPT Plus ($20/mo) if you only need one model. For heavy multi-agent workloads, the Ultimate tier at $39.9/mo offers 35M credits, which is competitive with direct API costs but requires careful tracking.
Setup time & first value
How long it actually takes to get something useful out of LobeHub — broken out by persona, not the marketing-page minute.
For a simple chat setup, you can be up in under 5 minutes with the cloud version (sign up, pick a model, start chatting). Setting up a custom agent with skills and MCP connectors might take 30–60 minutes, depending on your needs. Self-hosting takes longer—expect 1–2 hours for Docker deployment and configuration.
Switching to or from LobeHub
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From Poe: You can recreate your bots as Agents in LobeHub, bringing over your prompts and model preferences, gaining more control and self-hosting.
- ↗To another platform: You can export your conversation history and agent definitions (since they're stored as Markdown/Skills) and import them into a compatible system.
Integrations
Resources & Guides
- Documentationlobehub.com
Introduction
LobeHub is the next-generation agent harness designed to democratize AI power. Move beyond one-off, task-driven tools and build long-term agent teammates that grow with you in the world’s largest human–agent co-evolving network.
- Resourcelobehub.com
Blog
Updates from the LobeHub team and Community—product, engineering, and ecosystem.
- Resourcelobehub.com
Resources
Docs, guides, and Community resources to help you get the most out of LobeHub.
- Resourcelobehub.com
Changelog
Track new features, fixes, and improvements in LobeHub.
- Resourcegithub.com
GitHub
🤯 LobeHub is your Chief Agent Operator, organizing your agents into 7×24 operations by hiring, scheduling, and reporting on your entire AI team. - lobehub/lobehub
Tutorials & Learning
YouTube returned 6 videos for “LobeHub”, and we withheld 5: 5 could not be judged, because “LobeHub” is a single word that other videos use for other things. Showing the 1 we can prove is about LobeHub.
Official links
Popular in Agent Frameworks & Orchestration
Temporal AI
Temporal is the durable execution platform for AI agents and long-running workflows that survive crashes, retries, and abandoned sessions.
Frequently Asked Questions
Best-of guides
Used LobeHub? Help shape our editorial sentiment research.