---
title: "Langfuse September Update"
description: "Langfuse September Update: AI Filters, Experiment Runner SDK, Structured Outputs for Experiments, TypeScript SDK v4 GA, price drop, and more."
ogImage: /images/blog/2025-09-30-langfuse-september-update/september-update.png
tag: update
date: 2025/09/30
author: "Marc"
---
This month, we shipped our first AI feature, introduced a new high-level SDK to run experiments, added structured output support for prompt experiments, graduated our TypeScript SDK v4 to GA, and reduced the price of the Core plan. Here's what's new:
## Experiment Runner SDK
We've added a new high-level SDK abstraction that makes running experiments on datasets a breeze. It comes with automatic tracing, concurrent execution, and flexible evaluation built in. This is now the recommended way to run experiments on local or Langfuse-hosted datasets with our Python and JS/TS SDKs.
→ **[Learn more in the docs](/docs/evaluation/experiments/experiments-via-sdk#experiment-runner-sdk)**
## Natural Language Filtering for Traces
You can now filter your traces and observations using plain English. Just describe what you're looking for, and Langfuse will construct the right filters for you. You can then refine them manually if needed. This is our first AI feature in Langfuse, with more to follow!
→ **[See how it works](/changelog/2025-09-30-natural-language-filters)**
## Structured Output Support for Prompt Experiments
Ensure your LLM responses conform to a specific JSON schema in your Prompt Experiments. Enforcing a structured output makes it much easier to evaluate outputs consistently and process results programmatically.
→ **[Read the changelog](/changelog/2025-09-30-structured-output-experiments)**
## TypeScript SDK v4 (GA)
We’ve taken JS/TypeScript tracing up a notch: The Langfuse JS/TypeScript SDK v4 is out of beta as of September 2025! We’ve rebuilt this SDK on top of OpenTelemetry to improve the developer experience, make context management more robust, and allow for easy integrations with the JS/TS ecosystem (Provider SDKs, Vercel AI SDK, LangChain JS, Mastra, ...).
→ **[See the documentation](/docs/observability/sdk/overview)**
→ **[Upgrade now from v3 to v4](/docs/observability/sdk/upgrade-path/js-v3-to-v4)**
If you are on Python, all of this is already available in the current version of the Python SDK.
## More September Releases
- **[Mutable Score Configs:](/changelog/2025-09-29-score-config-updates)** You can now edit score configurations in Langfuse after they've been created.
- **[Core Plan now $29:](/pricing)** We've dropped the price of our Core plan by 50% from $59 to $29/month.
- **[Self-Serve Enterprise Cloud:](/pricing)** We've introduced a new self-serve Enterprise plan for teams that need advanced security, support, and scale.
## User Highlights
**Khan Academy:** uses Langfuse's public API to power Khanmigo AI, building a custom Golang client and linking internal tools directly to traces for debugging and analysis. Since rolling out in April 2024, adoption has expanded to 100+ users across product and infrastructure teams, enabling rapid iteration, shared visibility, and leadership insights without hosting their own tracing infrastructure.
→ **[Read the whole story](/users/khan-academy)**
---
## Agent Instructions
This page is part of the [Langfuse](https://langfuse.com) documentation, published as plain Markdown for AI agents. Every page is available as Markdown by appending `.md` to its URL, or by sending an `Accept: text/markdown` header. This page: `https://langfuse.com/blog/2025-09-30-langfuse-september-update.md`.
### Querying these docs
If the answer is not on this page, query the documentation instead of guessing:
- **Semantic search** across all Langfuse docs, returning an answer with the relevant pages and excerpts. Ask a specific, self-contained question:
```bash
curl -sG "https://langfuse.com/api/search-docs" --data-urlencode "query=How do I trace a LangGraph agent?"
```
- **Index of every page**: , with per-section indexes [llms-docs.txt](https://langfuse.com/llms-docs.txt), [llms-integrations.txt](https://langfuse.com/llms-integrations.txt), and [llms-self-hosting.txt](https://langfuse.com/llms-self-hosting.txt).
### Before writing Langfuse code
- **Install the [Langfuse Agent Skill](https://langfuse.com/docs/api-and-data-platform/features/agent-skill).** It encodes Langfuse's own best practices for instrumentation, prompt management, and evaluation, and materially improves results.
- **Read [What does a good trace look like?](https://langfuse.com/docs/observability/best-practices.md)** before instrumenting an application.
- **Verify endpoints, parameters, and response fields** against the [API reference](https://api.reference.langfuse.com) instead of inferring them from code examples.
- **Use the [Langfuse CLI](https://langfuse.com/docs/api-and-data-platform/features/cli)** (`npx langfuse-cli api `) to read or write traces, prompts, datasets, and scores from the terminal.
Found an error in these docs? Please open an issue at .