LangChain
LangChain is an MIT agent-engineering library. Rechecked 2026-08-17: github.com/langchain-ai/langchain had 144,367 stars, license MIT, description The agent engineering platform. The hosted commercial layer is LangSmith on langchain.com/pricing: Developer $0 / seat (up to 5k base traces / month, then pay-as-you-go) and Plus $39 / seat / month (up to 10k base traces). Enterprise is custom. The last LangSmith refresh also recorded Startup $1,400 / month and LangSmith Engine $2,000 / month on that page. LangServe-as-the-hero and a generic "context plus reasoning" essay are out.
The graph runtime lives on LangGraph. Compare Dify if you wanted a visual canvas instead of Python/JS.
Key Features
- OSS library: Python and JS packages for agents, tools, and model I/O. Treat current docs as the catalog, not a 2024 LangServe landing page.
- LangGraph for long loops: Multi-step agents moved to the LangGraph page. Do not describe LangChain as only LCEL chains.
- LangSmith for traces: Hosted evals, traces, and deploys are a separate bill.
- No framework subscription: The library is MIT. You pay models and, if you want it, LangSmith seats.
Use Cases
- Teams writing agents in Python or JS who will stay on GitHub until they need traces.
- Teams that will pay Plus $39 / seat for LangSmith after 5k free traces.
- People comparing LangChain vs Dify on code vs canvas.
Limitation: We did not re-verify LangServe as a current flagship product. Quote LangSmith dollars from the pricing page, not from this page's old "one-stop lifecycle" copy.
Pricing
Library is $0. Hosted rows from langchain.com/pricing on 2026-08-17.
| Path | Price | Notes |
|---|---|---|
| LangChain OSS | $0 | MIT. 144,367 stars. |
| LangSmith Developer | $0 / seat | Max 1 free seat in the compare copy we opened. 5k base traces / month. |
| LangSmith Plus | $39 / seat / month | 10k base traces / month, then usage. |
| Enterprise | Custom | Extra hosting and security. |
If a leftover blog still sells LangServe as the paid product, treat this table as the source.
Getting Started
- Install the Python or JS package from docs.langchain.com.
- Add LangSmith only after you see trace volume.
- Use LangGraph for durable agent loops.
- Recheck Dify if you wanted a canvas, not a library.
Frequently Asked Questions
Is LangChain $39?
No. $39 is LangSmith Plus per seat. The library is MIT.
Same as LangSmith?
LangChain is the OSS framework. LangSmith is the hosted trace/eval/deploy product.
Same as Dify?
No. Dify is a visual workspace. LangChain is code.
Alternatives
Tips
- Quote MIT library plus LangSmith $0 / $39.
- Do not paste LangServe as the 2026 product name.
- Recheck Startup / Engine rows on the pricing page if you are past Plus.
Conclusion
LangChain is a MIT agent library with an optional $39 LangSmith seat, not a LangServe slogan. Start at github.com/langchain-ai/langchain, then decide whether LangSmith already covers the traces you need.
Comments
No comments yet. Be the first to comment!
Related Tools
RAGFlow
ragflow.io
Open-source RAG plus agents. Rechecked: Apache-2.0, 88,659 GitHub stars. Cloud signup exists; no public dollar table. Not a generic document scanner.
LlamaIndex
www.llamaindex.ai
Document agents plus LlamaParse. Rechecked: Free $0 (10K credits), Starter $50, Pro $500. 1,000 credits = $1.25. OSS framework is MIT.
Coze
www.coze.com
ByteDance agent studio. Rechecked: public coze.com has no dollar table. Coze Studio on GitHub is Apache-2.0 with 21,459 stars. Not a free-for-anyone slogan.
Related Insights
After I Connected Obsidian to OpenClaw, It Started Helping Me Make Decisions
Once Obsidian stopped being just a place to store notes and started working with OpenClaw, it began helping me organize context, connect information, and improve real decisions.

Anthropic Subagent: The Multi-Agent Architecture Revolution
Deep dive into Anthropic multi-agent architecture design. Learn how Subagents break through context window limitations, achieve 90% performance improvements, and real-world applications in Claude Code.
Stop Cramming AI Assistants into Chat Boxes: Clawdbot Picked the Wrong Battlefield
Clawdbot is convenient, but putting it inside Slack or Discord was the wrong design choice from day one. Chat tools are not for operating tasks, and AI isn't for chatting.