Shipped · Observability

What lmnr-ai/lmnr shipped

The public repository of Laminar · github.com/lmnr-ai/lmnr

Written by FoxPlug from the repository's own releases; not affiliated with Laminar. 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/lmnr-ai/lmnr; Laminar 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 lmnr, free

Week of November 2, 2026

What shipped

Why it matters

This week brings a new release plus multiple dashboards per project and a cleaner span view overview, which change how people organize and read their traces and charts. It also adds Okta SSO with SCIM-managed membership for sign-in and access control, moves dashboard tooling into Next.js and the CLI, and fixes several crashes and validation mismatches in audited routes, agent version storage and trace ids.

Changelog entry

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

Post for X

V0.3.3 is out: multiple dashboards per project, span view overview inputs and outputs in their own formats, plus Okta SSO with SCIM-managed membership. Dashboard tools moved to the CLI, and audited routes and trace id validation got fixes.

Post for LinkedIn

Laminar v0.3.3 is out. This release adds multiple dashboards per project and renders span view overview inputs and outputs in their own formats, and it includes agent versions and faster ClickHouse migrations. Also this week: Okta SSO with SCIM-managed workspace membership, dashboard write and chart check tools moved from app-server to Next.js with dashboard write tools in lmnr-cli, and project_id now stored on agent records and document revisions with a composite foreign key. Fixes cover audited routes without dynamic segments, trace tags written through the API now updating trace tags, masked agent version text, and trace ids advertised as strings with a regex.

Week of October 26, 2026

What shipped

Why it matters

This week improves evals by making rows identifiable across runs, so results over subsets of datapoints compare correctly and repeated trials group under tasks. It also fixes dashboard and query behavior and eval realtime views, and makes spend estimates and upgrades clearer for people on trials. Self-hosted operators get a simpler way to wire Postgres roles for ClickHouse access.

Changelog entry

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

Post for LinkedIn

What shipped this week. Eval datapoints now carry a string task_id: rows are Trials, trials sharing a task_id in one run form a Task, and comparisons across runs over a subset of datapoints no longer match the wrong rows. Dashboards were fixed and the SQL to chart builder round-trip now fails loudly instead of silently ignoring unsupported WHERE predicates, with query formatting added. Custom columns on data, target and eval realtime views were fixed, and trial users can upgrade or switch plans from billing. Self-hosted deployments can reuse existing Postgres roles for ClickHouse SQL access, so the migration login no longer needs role creation rights.

Week of October 12, 2026

What shipped

Why it matters

This week shipped a new Starter trial path and moved agent versioning onto system-prompt versioning, so changes to prompts, tools and models are tracked per agent version. Self-hosted deployments also got fixes that matter day to day: the Compose frontend now carries the telemetry opt-out, and the Quickwit healthcheck no longer breaks docker compose up. Onboarding and signal notifications got smaller fixes, including a shorter signals step and single welcome email across tabs.

Changelog entry

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

Post for X

This week: a card-less 14-day Starter trial, agent versioning rebuilt on system-prompt versioning so prompt, tool and model changes are caught automatically, OpenCode V2 trace parsing, and fixes for Quickwit startup, signal scrolling and welcome emails.

Post for LinkedIn

This week's changes to Laminar: A card-less 14-day Starter trial is now available for new workspaces. Agent versioning checkpoints were rebuilt on top of system-prompt versioning, so changes to prompts, tools and models are caught automatically and can later be correlated with error rates per agent version. Self-hosted users got several fixes: the Quickwit healthcheck no longer depends on curl, so docker compose up works again after the Quickwit bump, the root telemetry opt-out now reaches each Compose frontend, and startup logs a clearer warning for older Quickwit deployments. Onboarding also got a shorter signals step and a welcome email that is sent only once across tabs.

Week of October 5, 2026

What shipped

Why it matters

This week brings a release with a compose change and a postgres schema migration that cannot be rolled back without a backup, so upgrade planning matters. It also lands role-based read restrictions and dataset CRUD, which change what data different people can see and manage. Alert emails and workspace usage views give teams more context when tracing events and reviewing usage.

Changelog entry

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

Post for X

V0.3.0 is out: Valkey in default Docker compose, a version bump, and a one-way postgres schema migration, so make a backup if you may roll back. This week also adds dataset CRUD, role-based read restrictions, alert email links straight to the event trace, and usage dedup savings.

Post for LinkedIn

V0.3.0 is out. The default Docker compose files now include Valkey, a bundled dependency was bumped, and the postgres schema migration is irreversible, so make a backup if you may want to roll back. Alongside the release, this week added dataset CRUD and role-based read restrictions, so read access and data management can be scoped more precisely. Alert emails now link straight to the signal event trace with the matching event selected. Workspace usage views show compression and dedup savings. Welcome and invitation emails were refreshed with updated copy and sender address, and email links now use the lmnr.ai domain.

Week of September 28, 2026

What shipped

Why it matters

The billing and pricing model changed substantially this week with Signals moving to usage-based pricing and the UI updated to reflect spend tracking. Template versioning went live for better trace content extraction. Multiple foundational pieces now support accurate credit accounting and clearer user visibility into costs.

Changelog entry

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

Post for X

This week: Signals pricing moves to pay-as-you-go with real-time usage tracking, user template versioning ships for trace extraction, and billing UI redesigned for clarity.

Post for LinkedIn

Ship week: Signals pricing model updated to pay-as-you-go with ClickHouse-backed usage tracking and clearer billing UI. User template versioning now enables per-version regex resolution for trace content. Setup CTA rotates agent icons, sidebar cards show spend against total credit, and blog rendering fixed for long-form content.

Week of September 21, 2026

What shipped

Why it matters

Custom gateway support through Responses API expands LLM profile flexibility for self-hosted deployments. PII masking improvements strengthen security for dual-mode projects. Refactoring span processing and fixing trace loading improve code maintainability and user experience.

Changelog entry

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

Post for X

Responses API now works with custom gateways. Fixed PII masking in dual-mode projects. Improved trace loading and responsive layouts. Better error handling for LLM profile setup.

Post for LinkedIn

This week we expanded LLM profile support with Responses API for custom gateways, strengthened PII protection through masking policies on dual-mode projects, and improved the trace viewing experience with responsive layouts and fixed loading behavior. We also refactored span processing into stage functions and added batch Redis operations for better performance.

Week of September 14, 2026

What shipped

Changelog entry

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

Post for X

Shipped dual-mode PII handling: store text once, apply masks per reader. Added dataset CRUD API. Fixed realtime agent_input masking, span preview extraction, and Quickwit memory pressure.

Post for LinkedIn

This week we shipped the final piece of trace-level access control: dual-mode PII handling stores canonical text with byte-range masks, letting different project members read appropriate versions of the same trace. We also added a dataset CRUD API for programmatic dataset management, fixed realtime masking for dual-mode projects, improved span preview extraction for reasoning models, and resolved a Quickwit memory issue at high traffic.

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.