Make.com Release Notes | JUN'26
What do the JUN’26 Make.com updates really mean for your team?
We break down what matters, what to act on, and where it impacts your operations.
https://help.make.com/2026
You get a more searchable, unified help experience and fewer broken builds through filter deprecations, plus operational quality upgrades like better scheduling, clearer HTTP/sleep module labeling, and improved webhook status visibility. For B2B automation teams, this strengthens maintainability and governance, but you must proactively review scenarios that use the deprecated data-store filters, test after UI/navigation changes, and update any integrations impacted by provider model deprecations. Net value: stability and faster troubleshooting, not new business logic.
https://help.make.com/unified-canvas-navigation
You can now tune how you navigate the Make.com canvas per device, including left vs right click panning and forcing an input mode when auto-detection is wrong. For your team, this reduces friction and builder time on laptops, tablets, and trackpads, which helps adoption and faster iteration. It is mainly a usability improvement, so prioritize it for builder productivity, not workflow governance or integration reliability.
https://help.make.com/new-version-of-api-endpoint-for-credential-requests-scenario-usage-tracking-connection-filtering-and-app-updates
You need to migrate to API v2 for credential requests by June 10, 2026, otherwise integrations may break; the parameter simplification should reduce integration friction but demands testing. Expanded scenario usage tracking and “my connections” filtering improve governance, cost attribution, and debugging across data stores and variables. New prompt modules add better time-context handling and token-based credit visibility, improving cost control and adoption, but you should update runbooks and monitoring to reflect the new tracking and credit behavior.
https://help.make.com/new-mms-options-app-updates
You get more control for Make Managed Services: create/manage child organizations, monitor data transfer and credit consumption, and reallocate credits via API. That supports scaling and governance across business units, but you must update your operational dashboards and automation ownership rules. App updates reduce friction: Claude Sonnet 4.6 in text prompt, Zoho multi-region connections, Typeform region selection, and a Zendesk OAuth refresh-token connection (legacy expires Apr 30, 2026). Plan migrations to avoid connection breakage.
https://help.make.com/unified-help-search-and-deprecation-of-filters
You get a more consistent in-app help experience while building and maintaining scenarios, which reduces time-to-fix and helps standardize how teams learn and review automations. The deprecation of four Data Store pattern filters affects future builds: you cannot select them in new scenarios, while existing ones keep running. Plan to refactor new logic using remaining filters or post-processing modules, and add regression checks so governance and adoption don’t suffer from silent workflow drift.
https://help.make.com/google-drive-personal-gmail-support-starred-apps-renamed-error-handlers-scheduling-improvements-claude-opus-48-and-webhook-detection-status-cards
Personal Gmail support for Google Drive widens automation reach beyond workspace-only data, enabling small teams and contractors to run scalable file workflows without extra tenant setup. App/module starring plus clearer “retry/skip” error terminology reduces builder friction and mistakes. Multi-run daily scheduling and better webhook detection status cards improve operational reliability and testability. Claude Opus 4.8 in Make can reduce manual step checking for multi-stage tasks, but you should govern prompts, data access, and approval points.
https://help.make.com/openai-gpt-54-nano-and-mini-now-available-in-make
You can now use OpenAI’s GPT-5.4 mini and nano in Make, giving you cheaper, faster model options for high-volume, multi-step, and agentic workflows. This helps you design tiered pipelines: use larger models for orchestration and mini or nano for repetitive subtasks like extraction, classification, routing, and triage. Expect better throughput and cost control, but tighten governance: add confidence checks, logging, and fallbacks to prevent silent quality drift across versions and tasks.
https://help.make.com/nested-if-else-and-merge-is-now-available
Nested if else and merge lets you keep multi-step business rules in one visible scenario flow, so your team can trace decisions without bouncing across router chains. You get clearer merge points, which improves maintainability and change review, especially for CRM, lead scoring, and service routing logic. Use it to group logically related conditions, but avoid overly deep branching that can become hard to test and govern without strong documentation and QA.
https://help.make.com/make-ai-agents-new-mcp-tools-are-now-available
You can now expand Make AI Agents by adding MCP tools, which lets your agents use the same external tool servers your team already uses in other AI assistants, without rebuilding HTTP glue. Strategically, this increases coverage for lead research, enrichment, and ticket handling and improves “one place to act” workflows. Operationally, you should govern tool permissions, validate outputs, and monitor latency and failure modes, since agent choice can change behavior when tool sets evolve.
https://help.make.com/introducing-mcp-toolboxes
MCP toolboxes let you package Make MCP servers into controlled AI tool sets, with selective tool access, per-toolbox authorization keys, and separate server URLs. For your automations, this is a governance upgrade: you can give different teams or clients only the scenarios they should use, while gaining visibility into which tools ran and reducing accidental cost or data exposure. Treat tool labels and keys as security boundaries; otherwise maintainability and drift still happen as scenarios evolve.
https://help.make.com/anthropic-claude-model-deprecations-on-june-15-2026
On June 15, 2026, Anthropic will deprecate Claude Sonnet 4 and Claude Opus 4 on the API, so any Make.com scenarios using those models will start failing at runtime with errors. You need to audit your existing Anthropic Claude modules, replace them with the supported model variants, and regression-test key workflows (lead scoring, summarization, support triage). Also re-check Anthropic pricing because replacements may change costs, and update monitoring so failures are caught fast.