SaaS Release — 2026-07-20 to 2026-08-02
Window: 2026-07-20 → 2026-08-02
Highlights
Section titled “Highlights”Memory for chats and agents
Section titled “Memory for chats and agents”Kindo memory is a governed product capability for chats and agents. Users can see when memory is enabled, manage remembered information, and opt out; admins can enable or disable memory for the organization and clear stored memory when needed. The memory backend is more durable across runs while preserving ownership and erasure controls.
Kimi K3 model support
Section titled “Kimi K3 model support”Kimi K3 is available in Kindo model selection surfaces and public APIs. Teams can use Moonshot AI’s Kimi K3 in agents and API automations alongside the rest of the supported model catalog.
Track webhook-triggered agent runs
Section titled “Track webhook-triggered agent runs”Direct webhook agent triggers return a Responses API run handle. Callers can use the handle to check status, poll output, and connect follow-up automation to the run they started.
Integrations
Section titled “Integrations”Broader MCP integration coverage
Section titled “Broader MCP integration coverage”Kindo added or expanded MCP tool coverage across several commonly used systems:
- ServiceNow: read record attachments and attach CSV outputs to incident tickets.
- Qualys: Container Security coverage, policy-compliance scan schedules, and clearer authentication errors.
- Saviynt: connector coverage for provisioning, job/task lookup, and identity-lifecycle workflows.
- Google SecOps: create and list reference lists.
- Linear and GitHub: initiative/project traversal, Dependabot and security alerts, and pull-request conflict workflows.
More reliable Microsoft and endpoint-security integrations
Section titled “More reliable Microsoft and endpoint-security integrations”Microsoft connection setup lets users choose the account to connect instead of silently reusing the browser’s existing Microsoft session. Microsoft Teams, SharePoint, CrowdStrike Falcon, and Guardicore flows return more accurate policy data or clearer error messages when something needs attention.
Improvements
Section titled “Improvements”Clearer spend and Command Center review
Section titled “Clearer spend and Command Center review”Credit displays include USD equivalents, making spend easier to understand at a glance. Command Center keeps the Personal/Org-Wide scope selector visible while scrolling and shows a clearer Personal-scope empty state instead of making the statistics section disappear.
Smoother search and interactive controls
Section titled “Smoother search and interactive controls”Search fields use consistent leading search icons, and chat sidebar search fields align with the rest of the app’s text-field styling. Rapid clicks on buttons, tabs, toggles, and similar controls no longer flash accidental text-selection highlights.
More accurate agent-run status and recovery
Section titled “More accurate agent-run status and recovery”Agent runs that fail abruptly move to a failed state instead of remaining stuck in progress. Durable child-agent execution is restored for long-running work, and model-stream failures surface as clearer chat errors instead of leaving the run opaque.
More reliable credit aggregation
Section titled “More reliable credit aggregation”Resolved a credit-aggregation overflow issue that could leave high-volume usage rows unaggregated and increase database load. Credit summaries are now more stable for organizations with large transaction volumes.