Experts is live · free · last verified 5 Nov 2026 · viability monitor
Experts
Experts.js links OpenAI Assistants together as tools to build multi-agent expert systems in JavaScript
If you're already on OpenAI's Assistants API in JavaScript, Experts.js fixes two concrete bugs for free: thread locking between tool calls and data loss when one agent summarizes another. It's a library, not a platform — MIT-licensed, small, and readable end to end. It won't design your multi-agent architecture for you, and model-agnostic or Python teams should look at AutoGen or crewAI instead.
Verdict generatedVerified 8d ago · liveness 56/100 · cite: rightaichoice.com/tools/experts
- JavaScript and Node.js developers building multi-agent systems on OpenAI
- Teams already using the OpenAI Assistants API heavily
- Prototyping agentic architectures without adopting a Python framework
- Engineers who want a small, readable library rather than a full orchestration platform
- Non-developers looking for a no-code or low-code AI tool
- Teams that want a managed SaaS product with a UI or dashboard
- Developers who need model-agnostic support across multiple LLM providers
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 Experts.js if you need model-agnostic orchestration across multiple LLM providers, a Python codebase, or a managed service with a UI — it is a JavaScript library bound to OpenAI's Assistants API.
Your OpenAI Assistants API usage is billed separately by OpenAI — Experts.js itself is free, but threads, file storage, and vector store search all draw on your OpenAI account.
Experts.js costs nothing to adopt — it is an MIT-licensed library — so the comparison that matters is engineering time, not licence fees. Compared with hosted orchestration platforms that bundle hosting and a visual builder, you save the subscription but pay in design and integration work; compared with AutoGen or CrewAI, the trade is JavaScript and OpenAI lock-in for a much smaller, readable code surface.
In short
Experts — Experts.js links OpenAI Assistants together as tools to build multi-agent expert systems in JavaScript. Best for JavaScript and Node.js developers building multi-agent systems on OpenAI, Teams already using the OpenAI Assistants API heavily, Prototyping agentic architectures without adopting a Python framework. Free to use.
What is Experts?
Experts — Experts.js links OpenAI Assistants together as tools to build multi-agent expert systems in JavaScript. Best for JavaScript and Node.js developers building multi-agent systems on OpenAI, Teams already using the OpenAI Assistants API heavily, Prototyping agentic architectures without adopting a Python framework. Free to use.
Experts.js is an MIT-licensed JavaScript/Node.js library from Ken Collins that strips the Run-object bookkeeping out of OpenAI's Assistants API. Instead of managing runs and threads by hand, you register assistants as tools, so a router or sales assistant can hand a question to a merchandising expert that itself calls an OpenSearch vector search. Each expert owns its domain and instructions, so one agent doesn't have to tote every tool definition and burn tokens on skills it never uses. The library targets two failure modes that show up in naive multi-agent setups. Every tool gets its own thread, so a tool can't share an assistant thread that's still waiting on tool outputs — that's the thread-locking bug sidestepped. Tools also control their own output, piping data straight into the parent's context rather than letting it get truncated or re-summarized hop by hop, what Collins calls the grapevine effect. Calling an expert is roughly one line: assistant.ask('question', threadId). The parent-to-child thread relationship is stored in OpenAI thread metadata, so a chat client only needs to hold one thread ID per conversation, typically a URL path parameter. The library leans on Assistants API capabilities Collins highlights in his writeup: vision for multi-modal threads, instructions up to 256,000 characters, up to 128 tools per assistant, and the Vector Store API for file search across up to 10,000 files per assistant. It's built for JavaScript and Node.js developers already on the Assistants API who want a small, readable layer rather than a full orchestration framework. If you need model-agnostic routing across providers or a Python ecosystem, look elsewhere — this is deliberately scoped to OpenAI's Assistants and to JavaScript.
Key features
- Links OpenAI Assistants together as tools to form multi-agent hierarchies
- Automatic per-tool thread management to avoid thread locking
- Tool-controlled output that pipes data to the parent context, fixing the grapevine effect
- Parent-to-child thread relationship stored in OpenAI thread metadata
- Single-call API: assistant.ask('question', threadId)
- One client-side thread ID per conversation, e.g. a URL path parameter
- Router assistant pattern with a panel of domain-owning experts
- Chaining tools such as an OpenSearch vector database search
- Vision support for multi-modal threads via the Assistants API
- Vector Store API integration for file search up to 10,000 files per assistant
- Instructions up to 256,000 characters per assistant
- Up to 128 tools per assistant
- Works with the OpenAI Assistants API and GPT-4o
- Open-source under the MIT license
- Built for JavaScript and Node.js
Behind the verdict
Where Experts.js earns its place is in the gap between "I have an Assistants API app" and "I have a multi-agent system." Collins built it to kill two specific bugs, and it does so without dragging in a framework. Thread locking goes away because every tool owns its own thread. The grapevine effect — where a sub-agent's real data gets summarized into mush by the time it reaches the parent — goes away because tools control their own output and pipe it into the parent's context. We'd reach for it when you've got a router pattern in mind: a company or sales assistant on top, domain experts underneath, maybe a search tool at the bottom hitting a vector database. In Collins's own example the merchandising expert owns an OpenSearch query tool and the sales assistant never has to know how to write one. That's a clean separation of instructions and tokens, and Experts.js makes the wiring short. The thread metadata trick matters more than it looks. One thread ID per conversation, stored client-side as a URL path parameter, and Experts.js finds or creates child threads as needed. If you're building a chat client, that's a lot less state to manage than hand-rolling runs. Where it bites: this is not a managed product and there's no UI, dashboard, or hosted orchestration. You're writing Node.js. It's also pinned to OpenAI's Assistants API and, at the time of Collins's writeup, GPT-4o — if you need model-agnostic routing across Anthropic, Google, or local models, this won't help you. Python teams standardized on AutoGen, crewAI, or MetaGPT are outside the target audience too; Collins says as much, and that JavaScript-first focus is the whole point of the library. There's also no third-party news or changelog signal on it recently, so treat the published feature set as the one
Models under the hood
as of 2026-11-09
Limitations
- Experts.js is a thin layer over OpenAI's Assistants API, which the source page describes as still in beta, so its capabilities are bounded by that API.
- It is a JavaScript/Node.js library that links OpenAI Assistants together as tools, with the author directing readers to the GitHub project documentation for a full breakdown of capabilities and options.
as of 2026-10-12
Researching Experts? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Is Experts still active?
Experts is live · free · last verified 5 Nov 2026 · viability monitor.
Viability score
How well maintained and how widely used is Experts? 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: November 2026
How we score →What do users say about Experts?
What people actually say about Experts — is it worth it?
Our community sentiment research written Jul 6, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
How we verified this
Every listed fact about Experts — pricing, features, links, status — is re-read from the vendor's own pages by our verification pipeline and each change is logged with a date; the last pass ran on 5 Nov 2026. Analysis is drafted with AI models and checked against our editorial standards before publishing; see how we verify.
Verification history
We have re-verified Experts 13 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.
- — changed SSL certificate
- — changed limitations
- — changed description, pricing tiers, who should skip it, features, who it suits, summary, our verdict
- — changed summary, who it suits, our verdict, features, pricing tiers, description, who should skip it, SSL certificate
- — changed limitations
- — changed switching out, hidden costs, features, pricing fit, description, switching in, setup time, our verdict, who should skip it
Showing the 6 most recent of 13 verification passes.
Free to cite with attribution — this page re-verifies continuously.
How much does Experts cost in 2026?
Experts: Free to use. Pricing last verified 5 Nov 2026.
Plans compared
Every published Experts tier, its price as the vendor publishes it, who it fits, and what it adds over the tier below. The 12-month cost projection follows.
| Plan | Price | Billing | Who it's for |
|---|---|---|---|
| Free | Free | — | JavaScript and Node.js developers who already build on the OpenAI Assistants API and want multi-agent composition without a licence fee.Adds: Starting tier — MIT-licensed open-source library at $0/mo, with your OpenAI Assistants API usage billed separately by OpenAI. |
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.
Where the pricing makes sense
The company stage and team size where Experts's pricing actually pencils out — and where peers do it cheaper.
Experts.js costs nothing to adopt — it is an MIT-licensed library — so the comparison that matters is engineering time, not licence fees. Compared with hosted orchestration platforms that bundle hosting and a visual builder, you save the subscription but pay in design and integration work; compared with AutoGen or CrewAI, the trade is JavaScript and OpenAI lock-in for a much smaller, readable code surface.
Has Experts' pricing changed?
No plan price for Experts has moved since we began logging in Aug 2026; the pricing page is re-read on every verification pass.
Who should pick Experts?
Pick Experts if you are JavaScript and Node.js developers building multi-agent systems on OpenAI, Teams already using the OpenAI Assistants API heavily, Prototyping agentic architectures without adopting a Python framework, Engineers who want a small, readable library rather than a full orchestration platform.
Look elsewhere if you are Non-developers looking for a no-code or low-code AI tool, Teams that want a managed SaaS product with a UI or dashboard, Developers who need model-agnostic support across multiple LLM providers, Python-first teams already standardized on AutoGen, crewAI, or MetaGPT.
Skip it if Skip Experts.js if you need model-agnostic orchestration across multiple LLM providers, a Python codebase, or a managed service with a UI — it is a JavaScript library bound to OpenAI's Assistants API.
GeneratedReal-world workflow fit
Concrete scenarios for the personas Experts actually fits — and what changes day-one when you adopt it.
Create a router assistant whose only tool is a merchandising expert, which in turn calls an OpenSearch vector database search tool; keep the single thread ID in a URL path parameter for the chat client.
Outcome: Product questions reach the database-backed expert without the router carrying OpenSearch instructions, and aggregate product data pipes back to the router's context instead of being summarized away.
Wire up a planner assistant that delegates sub-tasks to coding experts, using assistant.ask('question', threadId) as the only call surface and letting Experts.js find or create a thread per tool as needed.
Outcome: A working multi-agent prototype without hand-managing Run objects, and a codebase small enough to read end to end before committing to the architecture.
Chain an image tool such as Ideogram to a creative assistant and use Assistants API vision so generated artwork and feedback travel through the same thread as multi-modal messages.
Outcome: Generated images reach the parent assistant's context rather than being stranded in a child tool, so follow-up questions can reference the actual output.
Use cases
- Route customer inquiries from a sales assistant to a merchandising expert that can query a product database.
- Build a research assistant that delegates literature search, data extraction, and summarization to separate agents.
- Create a coding copilot where a planner breaks tasks into sub-tasks handled by dedicated coding experts.
- Run a multi-agent content pipeline: idea generator, outline specialist, writer, and editor as linked assistants.
- Assemble a financial analysis panel with agents for market news, portfolio risk, and regulatory compliance.
- Keep a personal creative assistant on-brand by chaining an image tool such as Ideogram with vision feedback.
- Build a company assistant that routes accounts, orders, and product questions to specialized experts.
Setup time & first value
How long it actually takes to get something useful out of Experts — broken out by persona, not the marketing-page minute.
For a JavaScript developer already comfortable with the OpenAI Assistants API: under an hour to install the library and get a single assistant answering via assistant.ask('question', threadId). A working router-plus-experts design with a database-backed tool is a day's work; a production multi-agent hierarchy with considered domain boundaries is a multi-day design exercise, per the author's own
Switching to or from Experts
How to bring data in from common predecessors and how to get it back out — written for the switcher, not the buyer.
- →From a hand-rolled Assistants API client: replace manual Run object management with assistant.ask('question', threadId) and register each downstream assistant as a tool.
- →From a single mega-assistant: split instructions and tools into domain-owning experts and keep a router assistant as the only entry point.
- ↗To Microsoft AutoGen: port the expert hierarchy into a Python agent framework if multi-provider support or Python becomes a requirement.
- ↗To CrewAI: rebuild the router-and-experts panel as roles and tasks if you want a higher-level, Python-first orchestration model.
Integrations
Resources & guides
Tutorials & Learning
YouTube returned 6 videos for “Experts”, and we withheld 6: 6 could not be judged, because “Experts” is a single word that other videos use for other things. We are showing none, because we could not prove any of them are about Experts.
Official links
What are the best Experts alternatives?
The closest alternatives to Experts we list are CrewAI, OpenAI Agents SDK and Designing Multiagent Systems. Ranked by direct product-type match, not category overlap.
| Alternative | Pricing | Starts at | Last verified |
|---|---|---|---|
| CrewAI | Freemium | Pricing not published | — |
| OpenAI Agents SDK | Free | Free | — |
| Designing Multiagent Systems | Paid | Pricing not published | — |
Alternatives to Experts
View allCrewAI
CrewAI is an enterprise multi-agent platform: visual builder plus a governed control plane for production agent workflows.
OpenAI Agents SDK
OpenAI Agents SDK: a free Python framework for building multi-agent workflows with handoffs, guardrails, and sandbox agents
Designing Multiagent Systems
Designing Multiagent Systems: a Python book on multi-agent LLM architecture, MCP, and picoagents
Tools that pair well with Experts
Common stack mates teams adopt alongside Experts, with the specific reason each pairing earns its keep.
CrewAI
CrewAI is an enterprise multi-agent platform: visual builder plus a governed control plane for production agent workflows.
OpenAI Agents SDK
OpenAI Agents SDK: a free Python framework for building multi-agent workflows with handoffs, guardrails, and sandbox agents
Designing Multiagent Systems
Designing Multiagent Systems: a Python book on multi-agent LLM architecture, MCP, and picoagents
Experts vs its closest rival
Experts's closest head-to-head is with Locus Robotics.
Our Experts vs Locus Robotics verdict written Jul 6, 2026 is more than 90 days old, so it is being re-reviewed against current data and is not shown.
Frequently asked questions
Categories
Best-of guides
Topics
For makers of Experts
- Get a Verified Listing for Experts ($39): a written decision within 1 working day.
- Free listing for another tool: Free reviews are decided in order, within 3 weeks of the review starting.
Used Experts? Help shape our editorial sentiment research.