Product updates
What shipped, and what it is for.
Stories about what changed in Rylvo and how it helps the people who build and run agents. Fixes and corrections are on the changelog.
October 2026
1 releaseSeptember 2026
9 releases- Improvement
A clearer path through the dashboard
Dedicated sections for Settings and key work areas make it easier to move between configuration, reviews, and releases.
Dashboard / Settings / Releases
- Improvement
Bot answer availability on the status page
The public status page now shows whether bots are answering, alongside platform health.
Status / Reliability
- Improvement
Faster website chat, with sources in the answer
The website assistant opens with a lighter client and can link to the business pages behind its answer.
Website Chat / Knowledge Base / Performance
- New feature
Enterprise identity, managed in your workspace
Enterprise owners can configure single sign-on and SCIM, verify company domains, and monitor SAML certificate expiry.
Enterprise / SSO / SCIM
- New feature
Typed decisions with TypeSafe Jev
Guardrails, routing, knowledge relevance, and other short decisions can use TypeSafe Jev, with trace visibility and an automatic fallback.
Jev / Guardrails / Knowledge Base
- New feature
From your website to a working assistant
Start with a website URL, build an assistant from its content, test the answers, and take it live from one guided flow.
Getting Started / Bots / Website Chat
- New feature
A leads inbox connected to your workflow
Captured leads now have a dedicated inbox, email delivery, and a direct path to CRM and automation tools.
Leads / CRM / Automation
- New feature
Client-ready results for agencies
Agency teams can share branded client results, preview a sample report, and send a monthly report automatically.
Agencies / Reporting / Clients
- New feature
Knowledge that respects access groups
Restrict a knowledge source to the end-user groups allowed to use it in answers.
Knowledge Base / Access Control / Enterprise
June 2026
10 releases- New feature
Build your workspace with Architect
Describe what you want to create, review a proposed plan, and apply approved changes through the same MCP tools available to external editors.
Architect / MCP / Workspace
- New feature
One Retention Pack across four workflows
Extend trace, Mission Control, audience, and conversation storage history together through one discounted add-on.
Retention / Billing / Governance
- New feature
Make recurring work and reporting automatic
Schedule tasks and client reports with flexible timing, AI analysis, and a clear view of what runs next.
Automation / Reports / Scheduling
- New feature
Separate test and live conversation storage
Playground conversations can flow to dedicated test destinations while live traffic keeps its production route.
Conversation Storage / Testing / Integrations
- New feature
Explore every agent turn in Traces
Follow model calls, tools, guardrails, knowledge retrieval, latency, and cost through one conversation view.
Traces / Observability / Analytics
- Improvement
More control over knowledge retrieval
Tune source retrieval with connection-level models, result limits, and relevance thresholds across test and live use.
Knowledge Base / Retrieval / Grounding
- New feature
Exercise agent safeguards with Red Team
Run automated adversarial scenarios and review how an agent responds across safety and edge-case categories.
Red Team / Safety / Testing
- New feature
Turn conversation patterns into agent improvements
Agent Evolution groups recurring signals from live traffic and proposes changes for an operator to review.
Agent Evolution / Quality / Review
- Improvement
Build and operate Rylvo through MCP
Connect compatible AI editors to a broad workspace tool surface for bots, knowledge, prompts, guardrails, and operations.
MCP / Tools / Integrations
- New feature
Compare prompt variants with real outcomes
Run prompt experiments, review results, and use evidence-backed suggestions to guide the next version.
Prompts / Experiments / Optimization
