The latest releases, features, and fixes shipped across the software industry — normalized into one feed. Browse all companies →
AI Search now supports the GLM-5.3 Flash model for text generation, which features a 1,048,576-token context window and runs on Workers AI. Users can configure this model for their AI Search instances.
Second release candidate for Appwrite 2.0.0 with fixes for build artifacts survival during upgrades from 1.9.x, versioned infrastructure migrations, optional account setup step, and new OAuth2 providers (Cloudflare and Resend). Pre-release version not recommended for production use.
Voice Call Screen Sharing is now stable with the beta alert removed and in-room call interface enabled for all calls. Fixed silent background loading of entire room history and attachments, and resolved a reload loop when opening links to older messages.
August 2026 update adds more control over how Copilot reasons, model selection, team-shared specialized agents, and code review workflows for Visual Studio users.
GitHub Copilot releases provide users with more control over how Copilot runs, including new team collaboration features in Slack and Teams, along with customization options across the app, CLI, and IDE.
Starting September 25, 2026, Google Sheets destination syncs require a primary identifier mapping to track row changes accurately. New syncs and edits to existing syncs must specify a unique source column as the key, either via the UI or API/Terraform with is_primary_identifier: true.
This release adds hook events for model switching, live streaming of subagent tool calls to Remote Control clients, spend limit tracking, and per-session prompt-cache metrics. It also includes numerous bug fixes for file operations, security issues, symlink handling, session management, and various CLI and UI improvements.
Updated @clerk/shared dependency to version 4.30.2 to include the latest patches and improvements.
Introduced langchain.mcp, a first-party adapter that connects any MCP server to LangChain agents via MCPAdapter, supporting multiple servers, flexible authentication, caching, and both legacy and modern protocol versions. Tools are automatically extracted and can be passed directly to create_agent for immediate use.
A hotfix release resolves an issue where a small percentage of DNS queries were failing across platforms.
Fixed a critical issue where the client could fail to connect or switch organizations due to invalid registration after version switches, and resolved DNS query failures affecting a small percentage of queries across platforms.
A new GA release is now available for the Linux Cloudflare One Client. This hotfix resolves an issue where a small but noticeable percentage of DNS queries were failing across platforms.
GitHub is making three separate upcoming changes to Copilot policies and billing to provide a consistent experience. Users should review the updates to understand potential impacts.
Fixed a race condition in the query-orchestrator's cold-cache renewal path that could cause issues during concurrent cache refresh operations.
Fixed an issue where domain-restricted credentials were unable to function within their own node. This patch restores the expected behavior for credential scoping in node execution.
Fixed a bug where domain-restricted credentials could not be used within their own associated node. This resolves credential validation issues for domain-specific authentication scenarios.
Cline Desktop now ships on Windows with auto-updating, improved shell handling, and inline image rendering for tool results. Major fixes include scheduled task persistence, better session search across full history, improved provider credential detection, OAuth sign-in error handling, and MCP server stability.
This release includes bug fixes and improvements to system reliability and stability.
You can now create and deploy agents directly from the Vercel dashboard with a builder that scaffolds the project, creates a private Git repository, and deploys it to your team. Configure the agent's identity, choose an AI model through AI Gateway, add a web chat or Slack interface, and connect tools and data from services like Linear, Notion, or custom MCP servers.
Hy4 Preview, a 770B parameter open-source Mixture-of-Experts model from Tencent, is now available on Vercel's AI Gateway. The model supports 1M token context windows and is optimized for coding, document analysis, and scientific reasoning, accessible via the AI SDK with no platform fees.
Neon added three new CLI commands to integrate Neon into coding agents like Cursor and Claude Code without manual configuration. New Neon backend guides and MCP stalled-query checks are now available.
Your Notion agent can now provide edit suggestions on your content, enabling AI-assisted improvements to documents and pages.
Vercel CLI now includes dedicated commands for managing DNS records, domains, and projects, enabling terminal-based configuration that was previously only available in the dashboard. New capabilities include updating DNS records, renewing domains, controlling project state and observability, and managing project members with JSON output support.
Durable Objects can now use up to ten distinct Dynamic Workers with in-flight requests, increased from four, enabling better concurrency for shared I/O contexts. Other Workers retain a limit of four distinct Dynamic Workers per request.
Cloudflare has released Z.ai's GLM-5.3 agentic coding model on Workers AI with substantial performance improvements over GLM-5.2, including 50% gains on Z.ai Code Bench and 6x improvements on Terminal Bench 3.0, while maintaining the same pricing of $1.40 per M input tokens and $4.40 per M output tokens.
GitHub has improved label management on issues to make it easier to organize and discover labels, especially helpful for repositories with extensive label lists. The feature includes suggested labels to help users find the right label more quickly.
Copilot code review now supports additional pull request types, including automatic reviews on bot-authored PRs and very large PRs, with improved resolution explanations for review decisions.
GitHub Classroom has been deprecated as of August 28, 2026, and its website, APIs, and services have been decommissioned in favor of selected partner solutions.
Initial alpha release of LangChain 1.4.0 with numerous improvements to MCP integration, middleware enhancements, agent creation, and tool handling. Includes fixes for content type conversion, docstrings, and multiple stability improvements across the framework.
Claude Code 2.1.248 adds restricted mode to disable built-in tools, experimental prompt cache TTL configuration, server-managed settings diagnostics, and cross-session messaging. The release includes numerous fixes for prompt cache misses, session management, permission handling, MCP servers, and terminal interaction issues.
Bug fixes for api-gateway pre-aggregation build job status tracking, query-orchestrator data source queue initialization, and server-core orchestrator memory management. Added concurrent websocket connection limits per user in cubestore.
Patch release of langchain-fireworks with bug fixes and improvements since version 1.6.0.
Grafana Loki 3.6.16 includes a new flag to ignore missing chunks during log deletion with line filters, a security update to etcd client, and a storage optimization to pre-compute SHA-256 for AWS PutObject operations.
Starting October 1, 2026, checks, workflow runs, and statuses will be governed by the same Actions retention setting that controls artifacts and logs, unifying retention policy management across all workflow data.
This release includes a flag to ignore missing chunks during log deletion with line filters, security updates for containerd, etcd client, and golang.org/x/mod dependencies, and a performance improvement for S3 PutObject operations by pre-computing SHA-256 hashes to avoid aws-chunked encoding.
This patch release reverts a prior release and includes fixes and improvements to structured tools, making StructuredTool JSON-serializable.
Introduces webhook data sources for Data Studio, preserves active workflow enrollments during record merges, and adds app object support to association workflow actions. These updates reduce custom integration work and improve automation reliability.
Patch release with bug fixes and updates. See pull requests #39894 and #39964 for specific changes.
When you block a user from your personal account or organization, you can now automatically close all their open issues, discussions, and pull requests in one action.
Release of langchain-anthropic version 1.7.0 with updates to thinking display mode, container support, and advisor tool beta header for the latest advisor API version.
The AI SDK harness layer now supports Cursor through an official adapter, allowing applications to switch between different coding agents without changing application code. The adapter uses the Agent Client Protocol to connect Cursor to the HarnessAgent interface.
HubSpot is permanently removing the ability to create new legacy private apps through the account UI. Users should migrate to Service Keys available in Developer Platform Projects 2026.09 or later for system-to-system integrations; existing legacy private apps remain unaffected.
First release candidate for Appwrite 2.0.0 self-hosted. Includes critical fixes for Postgres data persistence, cache key collisions, console schema migrations, and optional DocumentsDB/VectorsDB configuration. Not recommended for production.
The Deployments page now features redesigned filters that enable faster discovery of specific deployments. Users can apply suggested filters for common searches, type to find options, and use natural language to describe what they're looking for for automatic filter application.
This alpha release introduces a new -minimal-refresh planning option to refresh only resources with proposed changes, adds enhancements to logging and diagnostics, and fixes bugs in mathematical functions and ephemeral resource handling. It also includes experimental features for deferred actions, test cleanup, backend configuration in tests, and policy evaluation.
Development release adding Exasol adapter support with full materialization and macro bundle, parse-safe views for safer graph querying, new ClickHouse configs, and improvements to reference tracking, logging, and security including symlink extraction fixes.
Fixed issues with glibc and libatomic in distroless runners image by staging them as matched pairs to ensure proper runtime compatibility.
Vercel introduces Workflow SDK, a new durable execution framework that lets developers write workflows as normal sequential code without complex DAG configuration or infrastructure management. The open-source, spec-based system uses only standard infrastructure like databases and queues, with client-side execution and automatic versioning to handle code changes without breaking in-flight runs.
SendGrid now offers automatic SSL certificate provisioning for branded links, eliminating the need for external certificate management. A new auto_ssl parameter in the Link Branding API enables this feature, with certificates provisioned and hosted by SendGrid alongside traditional link branding DNS records.
Plugin marketplaces in enterprise-managed settings can now be configured for automatic updates by setting autoUpdate: true on extraKnownMarketplaces entries. Supported clients automatically check and update these marketplaces.
This release adds the SendFeedback tool for drafting error reports, enhances API cost profiling via /claude-api cost-optimize, improves plugin marketplace hardening, and fixes numerous bugs affecting keyboard shortcuts, background sessions, shell commands, and terminal integration across various keyboard layouts and edge cases.
Inclusion AI's Ling 3.0 Flash Fin, a finance-focused LLM with 256K token context and 32K output tokens, is now available on Vercel's AI Gateway at no cost through September 25. The model supports reasoning and function calling for financial research and multi-step analysis tasks.
Fixes a bug in cubestore-driver where queries would fail on write EPIPE errors, improves error message readability for over-limit scenarios, repairs the release process in lerna-publish.sh, and adds support for pushing UNION queries down to the data source for better performance.
GitHub is rolling out enforcement of a default model policy for generally available Copilot models across Copilot Business and Enterprise plans. This policy was previously announced in July and is now being gradually enforced.
Bug fix release addressing user ban expiration handling, client type compatibility, session data validation, database migration issues, and URL validation improvements across multiple packages including core auth, OAuth provider, and database adapters.
Snyk CLI v1.1307.0 introduces secrets test exclusions, container image provenance attestations by default, offline .NET/NuGet analysis, experimental agent command space for AI coding agents, and multiple reliability and performance fixes including support for pnpm v11, Gradle 7.4–8.2 crash fixes, and improved handling of special characters in exclusion rules.
Terraform v1.16.0 introduces provider-specific state preservation, new terraform_data store blocks for ephemeral values, support for import blocks in modules, nested blocks as computed values, and numerous enhancements including machine-readable output for state and workspace commands, Mermaid graph formatting, and expanded action configuration capabilities.
Fixed a bug where query timezones were not validated against the IANA timezone database, which could cause incorrect results. Now queries with invalid timezone values are properly validated.
8 more changes from before the last 3 days. Sign up to read the full archive.
Sign up freeGitHub or email — no card needed.