Switch Trust changelog

• Guardrail savings - An agent’s Improvements tab shows the cost avoided by prompts that guardrails blocked, and the potential additional savings from changing alert-only policies to block. See Improvements
• Past analyses - An agent’s Improvements tab lets you browse its earlier analyses, each with its status: queued, running, finished, or failed
• Filter guardrails charts by time range - On an agent’s Sessions tab, show the guardrails charts for the last 24 hours, 7 days, or 30 days
• Promotion codes at checkout - Enter a promotion code when you upgrade your plan

• Model pricing updates - Cost tracking adds GPT-6.1 Terra, corrects cached-input pricing for Gemini 4 Argon, and includes cache-write pricing for current OpenAI models. See Model pricing

• Changing an AWS role - Saving a different IAM role for an AWS data source replaces the previous role, and a successful save clears an earlier connection error. See AWS

• Platform hardening - Stronger separation between workspaces, a stricter content security policy for the web interface, and tighter validation of service account keys

• Ingest-only API keys - The Ingest only role lets a key send data to Switch Trust and read guardrail policies, but not read your data. Sensor setup flows select it for you. Use it for sensors such as the SDKs, OpenTelemetry, the Claude Code plugin, and the GitHub and GitLab discovery actions, so a leaked key can’t expose conversations or inventory. See API keys
• Filter evaluations by type - Filter the evaluation list to show built-in or custom evaluations

• A clearer way to add agents - The Agents page and the dialog for adding agents list every way to add agents, recommended routes first, each with a line saying what it does. The routes cover the Claude Code plugin, Switch Rooms, OpenTelemetry, scanning your repositories, evaluating an agent, and monitoring and protecting one
• Faster answers when an agent endpoint is wrong - If an agent endpoint’s host doesn’t exist, such as an expired tunnel, the evaluation run stops right away and says so, instead of retrying first
• Organization names in any language - Organization names can include non-ASCII characters
• Sessions from every source - An agent’s Sessions tab shows activity from OpenTelemetry or the Claude Code plugin even when the guardrails SDK isn’t installed, with a prompt to set up guardrails protection alongside it
• Users page shows active members - The Users page shows how many members are active alongside the total
• PII and secret detection catches more - Guardrails catch PII and secrets written with JSON \uXXXX escapes in messages and tool results, and PII in long messages that was sometimes missed or only partly redacted. The Regex detector also matches escaped message text, so a <script> pattern now matches \u003cscript\u003e. See Guardrail policies

• Links survive signing in - A Switch Trust link opened while you’re signed out takes you to that page after you sign in
• Improvement details stay open - The detail panel on an agent’s Improvements tab stays open when you switch between optimizations
• Platform reliability - Improvements to accessibility, the issue occurrence list, and API requests

• Platform hardening - Tighter network and permission checks across the platform

• Read an agent’s spending cap through the API - Fetch the monthly spending cap set for an agent, if it has one. See API reference

• Platform reliability - Improvements to evaluation scheduling, finding tracking, OpenTelemetry ingestion, and API paging

• Claude Code and Switch Rooms as data sources - The Data sources page has setup steps for the Switch Trust Claude Code plugin, which reports Claude Code usage and cost, and for connecting your Switch Rooms workspaces. Switch Rooms activity is grouped under one agent with a Switch Rooms activity badge, and each room appears as its own session
• API identifiers in organization settings - Organization settings show your Tenant ID and Current workspace ID, each with a copy button, so you have the values that Switch Trust API request paths need. See Organization
• Documentation and support from the top bar - Links to the documentation and to support are in the top bar on every page

• New sign-in address - Switch Trust is now at app.switchagents.ai, as part of the move to the Switch AI name
• Evaluation failures say what went wrong - A failed evaluation run gives a specific cause, such as rejected credentials, a provider rate limit or quota, a timeout, or an unreachable endpoint, and names the endpoint involved, instead of a generic error. See Agent evaluation results
• Evaluation schedules on the Free plan - On the Free plan, evaluations run on a weekly or monthly schedule. Running an evaluation now, or on a daily schedule, is available on paid plans. See Plans and billing
• Spending cap summary as a bar chart - The spending cap summary shows how many agents are below, near, and over their cap, and how many have no cap set. Caps keep four decimal places, so a small cap is stored exactly as you enter it. See Spending and caps
• Fewer false PII detections - PII detection in guardrails and evaluations no longer flags hashes, timestamps, dates, or five-digit numbers as personal data, and redaction no longer alters a longer number that only partly matches a PII pattern. See Guardrail policies
• Pricing for new models - Cost tracking includes pricing for Claude Sonnet 5.5, Claude Haiku 5.5, GPT-6.1 Sol, and Gemini 4 Argon. See Model pricing
• Faster inventory table - The inventory table loads faster, especially for organizations with a large inventory
• Clear not-found pages - A link to an agent or guardrail policy that no longer exists, or to a page that doesn’t exist, opens a page that says so

• Assets page opens reliably - The Assets page loads when you open it from a link or refresh it
• Setup snippets show current values - The GitHub data source snippet shows the address of your Switch Trust app, and the GitLab snippet uses the current version
• Platform reliability - Improvements to billing, data source scans, and evaluation runs

• Platform hardening - Tighter authentication and request validation across the platform

⚠️ Action required

• Guardrails need an Editor API key - Guardrails now refuse a request sent with a Viewer API key, with a 403 response that names the missing permission, so a misconfigured key is caught immediately. Previously, the request reached the model but nothing was recorded. If your agent uses a Viewer key, replace it with a key that has the Editor role or higher. See API keys
• Stricter API filters - List endpoints now return an error for a filter field they don’t recognize, rather than ignoring it, so a mistyped filter no longer returns unfiltered results. Operations that the API reference listed but the API never served have been removed. If you call the API directly, check that your requests use only the filter fields in the API reference

• Plans and billing - Choose the Free, Standard, or Enterprise plan from Plans & billing in Settings. The page shows how many managed agents your plan allows and how many you’re using, and lets you pay by card and see your invoices. If a payment fails, a banner links you to update your payment details. See Plans and billing
• Retry and loop waste - An agent’s Improvements tab flags spend lost to repeated tool calls and resubmitted prompts within a session, with an estimated monthly saving. See Improvements

• Issues scoped to the agent you’re viewing - Opening an issue from an agent’s Issues tab shows only that agent’s instances of it. The main Issues page still lists every affected asset
• Custom evaluations checked when you save - A custom evaluation’s configuration is validated when you save it, so a mistake shows up immediately rather than when the evaluation runs

• Platform reliability - Improvements to billing and agent analysis

• Flint AI is now Switch Trust - The product is renamed Switch Trust throughout the interface. Your agents, findings, and settings carry over unchanged
• Resize detail panels - Drag the edge of a detail panel to make it wider or narrower, so you can size it to what you are viewing
• Data source connection status reflects the last test - A data source’s connection status now follows its most recent test connection across GitHub, GitLab, AWS, and Google Cloud, so a connected or disconnected state matches the last check

• AWS and GCP as data sources - Connect an AWS account or a Google Cloud project in Settings, then Data sources, and scan it for the AI resources deployed there, alongside your GitHub and GitLab repositories. Switch Trust connects read-only, through an assumed IAM role for AWS and Workload Identity Federation for GCP, so there are no long-lived keys to store

• Long text stays readable in detail panels - A detail panel expands long text in place instead of cutting it off, so you can read the whole value without it being truncated

• Platform hardening - Internal-only service endpoints are separated onto a mesh-internal port so they are not reachable from outside the platform network

• Clearer required fields when creating a guardrail policy - The policy form marks which fields are required before you can create a policy, and when topic control is enabled it requires the agent objective and guidelines so the detector has something to work with

• Inventory lists page reliably - Paging through an inventory list no longer shows the same row twice or skips one, so what you see across pages is complete and stable
• Prompts and responses display reliably - Prompt and response text is captured and retrieved correctly, so it shows up where you expect instead of coming back empty

• See which agents are managed - The agents list flags which agents are managed and shows the count, so you can tell at a glance what is under management
• Failed LLM interactions show as errors - An interaction that errors is now surfaced as an error instead of appearing empty, so a failed call is visible where the interaction is

• Tool-result redaction now covers Anthropic - Guardrails redact PII and secrets in the tool results an agent receives on Anthropic, extending the coverage already in place for OpenAI and Gemini

• Sort agents and inventory by issue count - The agents table and the inventory tables sort by number of issues, so whatever needs attention rises to the top
• See where each tool came from - Every tool in your inventory records the source it was discovered from, shown as a filterable column in the tools table. Sources are derived from your agents’ real interactions during analysis and refresh when a repository is scanned again
• Build policy tool lists from your inventory - When you set a guardrail policy’s tool allow or block list, pick from the tools already discovered in your inventory instead of typing names by hand
• Assign and manage member roles - Choose a member’s role when you invite them, change a member’s role afterward, and see the role attached to each API key. The interface shows only the actions your own permissions allow

• Invitations land you in the right organization - Accepting an invitation moves you into the organization that invited you

• Tool-result redaction now covers Gemini - Guardrails redact PII and secrets in the tool results an agent receives on Gemini, matching the coverage already in place for OpenAI

• See what a tool can do - Tools in your inventory carry capability tags such as network access, filesystem writes, and shell execution, shown in the tools table and on each tool’s detail panel. Each capability records how it was determined, so a resolved call site is distinguishable from a match against a catalog of known tools. Tags fill in as your repositories are scanned again
• Cached tokens are priced separately - Cost now prices cached tokens apart from fresh ones, so prompt caching shows up as the saving it is instead of disappearing into the input count. Session and turn views break out cache hits and cache stores, and you can set your own cache read and write rates alongside input and output on the Model pricing page
• Resend a pending invitation - Through the API, admins can send the email for a pending invitation again instead of revoking it and starting over. An expired, accepted, or revoked invitation still has to be recreated

• Input token counts exclude cached reads - OpenAI and Gemini report cached tokens inside the input count. Those are now subtracted and reported on their own, so input figures for agents on those providers read lower than before. Activity already recorded keeps the counts it was stored with

• Account setup no longer needs admin permissions - Accepting the license agreement and completing the qualification step are open to every member, so people who are not admins can finish setting up their account
• Platform reliability - Guardrail services start reliably after a cold start, and requests are handled consistently across model providers

• Role changes respect the hierarchy - Nobody can grant a role above their own or act on a member ranked above them, and an organization can never be left without an admin. A blocked attempt returns the reason
• Platform hardening - Tightened the runtime configuration of guardrail services and widened vulnerability scanning of platform images

• Switch between organizations - If you belong to more than one organization, Switch organization in the account menu moves you between them without signing out
• Admin view of every API key in your organization - Through the API, admins can list all keys in the organization, personal and organization-owned, see who owns each one, and revoke any of them
• List the agents Switch Trust is tracking - A new endpoint returns the agents that have sent guardrail traffic or finished an evaluation, with the date each was first seen. The list is cumulative, so an agent stays on it after it goes quiet
• Cost for the latest Claude and GPT-5 models - Default pricing adds four current Claude models and the GPT-5 mini and nano tiers. Cost is calculated when you view it, so spend already recorded against these models appears too, not only new activity

• GitHub and GitLab are reported separately - Inventory records which platform each agent was found on, so results from the two can be told apart. Existing GitHub results are unchanged, and the new value appears once your scanner is updated

• GitLab as a data source - Connect GitLab in Settings, then Data sources, and scan its repositories for agents, alongside GitHub
• Sessions on the agent detail page - The Sessions tab groups an agent’s activity into sessions, and a detail panel breaks each session down turn by turn
• Event detail inside a turn - Open a turn in Live or on an agent to see the events it contains, including tool calls and the content tools returned
• Interactive asset graph - The graph on an agent’s Overview and Assets tabs expands to show what the agent connects to, and opens fullscreen

• Readable large numbers - Token counts and costs are abbreviated in tables and charts, so long numbers no longer stretch a column
• Consistent model health score - The score reads the same way in the models table and in the model detail panel
• Detail panels close from the main view - Selecting the main content closes an open detail panel

• Tool results in Live events - Tool result content displays instead of appearing empty
• Asset table dates - Last seen and created dates display correctly in the models, MCP servers, and tools tables

• API keys are revoked when a member leaves - Removing someone from your organization now revokes the API keys they created, so access ends with the membership
• API keys inherit the role of whoever creates them - A new key takes the creator’s role and can no longer be issued with a higher one
• Blocked content is stored redacted - When a guardrail blocks a message, the retained record no longer holds the original content
• Tool results are covered by redaction - Guardrails read the tool results an agent receives, so redaction applies to that content too

• Cost tracking for agents - Agent detail pages chart cost by month, cost per session, and burn rate, and the agent table adds Cost (MTD), $/session, and Cap usage columns
• Agent spend caps - Set a monthly spend cap on an agent, with a cap-aware burn rate and a days-to-cap forecast, so teams can track and manage agent spend
• Cost data page - Spend and per-model pricing in one place

• Automatic re-authentication - Sessions prompt a new sign-in when authentication expires, keeping access secure

• Platform reliability - Service availability, update handling, and API request validation improved across the platform

• Platform hardening - Tightened service security configuration and updated dependencies

• Organization member removal - Remove a member from your organization in the interface or through the API
• Identity lookup from API keys - A new endpoint returns the identity behind an API key, so you can attribute activity in your own logging and workflows
• API reference - Platform endpoints are documented in the API reference

• Guardrail coverage stays in place - An organization always keeps at least one active policy
• Guardrails availability during updates - Guardrails keep evaluating while the platform updates

• Guardrails under concurrent load - Detectors evaluate reliably when many requests arrive at once
• Connected assets on the agent detail page - Connected assets now display consistently
• Gemini requests without a role field - Improved compatibility for requests that omit a role
• Inventory finding severity - Inventory findings display with their intended informational severity
• Organization creation - Creating an organization produces a single organization, even if the request is retried
• Invitations for existing members - Inviting someone who already belongs to the organization returns a clear conflict response

• Sign-in and organization setup - Sign in with GitHub, Google, or email, then create organizations and workspaces, invite teammates, and issue organization-scoped API keys
• GitHub-based agent discovery - Find the agents in your repositories with a GitHub Action. See Discover your agents
• Central agent inventory - See the models, prompts, tools, MCP servers, and sub-agents behind every agent in one place
• LLM-based scanning - Scan agent code and MCP servers for issues
• Guardrails onboarding through the Switch Trust Python SDK - Connect running agents to the guardrails proxy. See Monitor and protect your agents
• Policy-based guardrails - Detectors for jailbreaks, toxicity, PII, secrets, PHI, and custom regular expressions, plus tool allow and block lists
• Configurable enforcement actions - Block, redact, or alert when a policy matches