Shipped · Observability

What langwatch/langwatch shipped

The public repository of Langwatch · github.com/langwatch/langwatch

Written by FoxPlug from the repository's own releases; not affiliated with Langwatch. Paste your own repository in the box below and FoxPlug writes the same weekly update for your product. This page is an automatic summary of the releases, pull requests and commits in github.com/langwatch/langwatch; Langwatch did not write it and does not use or endorse FoxPlug. Every line links to the change it describes.

Get a weekly update like this for your product, free

Or Use it as a GitHub Action

See the update FoxPlug writes for langwatch, free

Week of November 2, 2026

What shipped

Why it matters

This week's changes fix real dead ends for people running self-hosted installs, from licence renewal to SSO sign-in behind a proxy to seat limits. Billing and seats also got a lot of work: any member can start a subscription, seat counts follow what is in use, and the Usage page now shows how the month built up. Alongside that, API keys can be reopened and rotated, and version 3.23.0 brought agent and evaluator versioning with deploy labels.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: seats are held 30 days after a member is removed, over-limit organizations hold back SSO sign-ins, the Usage page was redesigned, any member can start a paid subscription, and API keys can be reopened and rotated. Version 3.23.0 also shipped agent and evaluator version

Post for LinkedIn

Several fixes landed this week for self-hosted installs, including licence renewal and replace, activation, and SSO sign-in behind a proxy. Seats changed too: removing a member holds their seat for 30 days and Stripe lowers the count from the next cycle, organizations over their seat limit now hold back SSO sign-ins, and any member can start a paid subscription. The Usage page was redesigned with a plan strip, a month-by-day view and a per-project split. API keys can be reopened and rotated. Version 3.23.0 shipped agent and evaluator versioning with deploy labels, plus experiment and dependency fixes.

Week of October 26, 2026

What shipped

Why it matters

This week two client packages moved to new major and minor versions, so dashboard and reports code written against the old API needs updating. Alongside that, the gateway gained a route for a budgets-and-tracing-wrapped API, and a set of fixes landed for trace reads, queue scheduling, SSO sign-in messages and Bedrock caching.

Changelog entry

Week of October 19, 2026

What shipped

Why it matters

This week closes several reliability and security gaps that showed up in real deployments: failing monitors and timed-out evaluations no longer run unchecked, forwarded headers are only trusted from known proxies, and API requests now record their credential type and client. Self-hosted and cloud setups also gain more deployment options through web-identity credentials, a new model provider and a nightly suite that reports in one place.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: monitors pause themselves when evaluations mostly fail, timed-out evaluation calls are no longer retried, API logs record credential type and client IP, Doubleword is a built-in provider, and a single nightly suite reports on one dashboard.

Post for LinkedIn

This week's changes: online evaluation monitors now pause themselves when most evaluations fail, so a hanging endpoint stops spending on failed runs. Timed-out evaluation calls are no longer retried. Each API request records which credential type and client it used, and legacy project keys are limited to ingestion. Client address headers are read only from configured proxies. The NLP Lambda clients and per-organisation S3 accounts accept web-identity credentials instead of static AWS keys. Doubleword is now a built-in model provider, Settings has an Instant Evals column for choosing the decision model, and the new nightly suite puts every result on one pinned dashboard issue.

Week of October 12, 2026

What shipped

Changelog entry

Week of October 5, 2026

What shipped

Why it matters

This week tightens the platform people already run: analytics dashboards hold up under high-volume projects, traces and experiments give clearer results, and self-hosted installs gained secret rotation and safer upgrades. Governance, voice in the gateway, and the new judge model open up new ways to read every trace and to serve consumer voice apps. Several fixes came straight from real installations and customer reports.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

This week: Instant Evals as the model for every LLM judge, analytics panels that stay under the memory limit, voice support in the gateway, an aggregate governance project, OTLP bodies decoded by magic bytes, and CREDENTIALS_SECRET rotation for self-hosted installs.

Post for LinkedIn

This week's work on LangWatch covers several areas people asked about. Instant Evals are now a model choice on every LLM-as-a-judge evaluator. Analytics dashboards stay under the ClickHouse memory limit and show a compact panel error rather than failing. The gateway gained voice support for consumer apps, with streaming audio, provider tokens, a live broker and realtime metering. Governance added an aggregate project that reads member projects through shared grants, and anomaly alerts now deliver to webhook endpoints with raw body and sha256 signing. OTLP bodies are decoded by magic bytes, so zstd and gzip payloads arrive correctly. Self-hosted installs can rotate CREDENTIALS_SECRET, scrape worker metrics, and drop two unused chart settings.

Week of September 28, 2026

What shipped

Why it matters

Two releases shipped this week addressing self-hosted security, SSO, and seat licensing. The strict layout branch landed several features including developer seats and improved notifications. Multiple hardening fixes for code isolation, authentication, and data privacy ensure self-hosted installs are more secure by default.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

3.20.1 and 3.20.0 shipped: stricter self-hosted defaults, SSO fixes, developer seats for enterprise teams, and improved notifications. Traces now display sub-penny costs correctly.

Post for LinkedIn

LangWatch 3.20.0 and 3.20.1 are out. Self-hosted installs now disable third-party calls by default and restrict self-sign-up via environment variable. A new developer seat lets enterprises add members to personal projects without consuming Full seats. Langy's notifications use Web Push. Instant Evals require opt-in before sending data outside. Traces display costs below a tenth of a cent.

Week of September 21, 2026

What shipped

Why it matters

3.18.0 brings self-hosted deployments to Kubernetes with a Helm chart, adds support for images and files in datasets, and enables self-hosted instances to use LangWatch-hosted services with metered usage. Teams using LiteLLM as their gateway can now trace all calls without code changes. These updates address deployment, extensibility, and integration gaps.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

LangWatch 3.18.0: Helm chart for self-hosted K8s, images and files in datasets, LiteLLM Proxy integration, connected self-hosted with metered hosted services.

Post for LinkedIn

LangWatch 3.18.0 is out. New Helm chart for self-hosted Kubernetes deployments. Datasets now support images and files. LiteLLM Proxy integration lets teams trace calls through their internal gateway without code changes. Connected self-hosted installs can use LangWatch-hosted services with metered usage against the license.

Week of September 14, 2026

What shipped

Why it matters

Instant Evals shipped as a complete feature this week: runs, CLI, judgment functions, billing, and metrics. LangWatchQL now has a unified query door that any API key can use. These changes let users evaluate bulk text through the platform and access analytics programmatically without building custom integration.

Changelog entry

Example posts FoxPlug drafted from these changes. Not written or posted by the project.

Post for X

Instant Evals are live: run bulk judgments through LangWatchQL, use the CLI without writing queries, and only pay for what you evaluate. One query door for any API key. Dashboard charts now run in sandboxed frames.

Post for LinkedIn

Instant Evals shipped this week with three layers: LangWatchQL now has eval() functions for inline judgments, runs execute bulk evaluation jobs with progress tracking, and the CLI makes it accessible without SQL. Billing integrates with Stripe on the standard cost spine. Separately, any LangWatch API key now has one query door into LangWatchQL—POST /api/v1/query runs read-only SELECT across all readable projects. Custom dashboard charts moved to sandboxed frames. Sign-in gained one-click social auth.

Weeks with too little public activity are left out rather than filled in. Last updated 2026-11-10.

Is this your repo? Ask us to remove this page.