> For the complete documentation index, see [llms.txt](https://docs.visdom.virtuslab.com/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.visdom.virtuslab.com/integrations.md). # Integrations ## Source control and issue tracking * Reading ticket content, comments, attachments, linked issues, change history and metadata (priority, component, assignment, labels, custom fields) through the tracker's REST API, from a service account with minimal permissions. * Configurable triggers: ticket created, status changed, label applied, assigned to a queue. * Writing feedback comments: processing status, pull request link, failure reason. * Creating branches, commits and pull requests through the source control REST API; setting title, description, reviewers and the link back to the ticket; marking the pull request as AI-generated. * Pull request webhooks feeding Code Review and the metrics. * The source control permission model is respected: agents act through dedicated service accounts with permissions narrowed per project and repository. The platform has a production GitHub integration; adapters for other systems are delivered as part of a deployment. The adapter architecture is built for this: an adapter implements inbound events and outbound operations against a common contract. ## Models: bring your own LLM * The driver layer separates the agent from the model vendor: a native Anthropic driver and an **OpenAI-compatible driver with a configurable endpoint**, which means full compatibility with LiteLLM-class proxies and any OpenAI-compatible gateway you operate. * Model choice and change without rebuilding anything: the model is an attribute of an agent definition, and versions are pinned explicitly, never "latest". * Model allowlists enforced by policy: using a model outside the list is blocked and logged. * A fully internet-isolated scenario: models hosted on your infrastructure (vLLM, Triton, open-weight models) behind the same gateway. The platform is functional with no outbound connectivity at all. * Full call telemetry: model, token counts, status, cost and duration, per call, per flow and per team. ## Identity: OIDC, SSO, MFA, directory * User authentication over **OIDC** (native, multi-IdP); OAuth2 authorisation; sessions in secure cookies. * An identity broker for SSO, MFA enforcement and **role and group synchronisation with Active Directory or Entra ID** (LDAP and SCIM federation). Direct integration with an existing OIDC-compliant provider is equally possible; SAML is supported through a SAML-to-OIDC bridge. * Unique user identifiers (no group accounts); unused accounts blocked rather than deleted. * RBAC with granular roles: global administrator, project owner, developer, read-only, plus auditor roles, with permissions per project and repository in the Control Plane. ## Secrets * Integration with your secret store as the source of truth: the platform pulls secrets just in time and stores only references. Secrets at rest inside the platform are encrypted with AES-256-GCM using a key ring, with KMS and BYOK support. * Secrets never reach the agent environment in clear text: substitution happens at the proxy layer. ## SIEM and observability * Audit event export to your SIEM: Syslog, structured JSON, Splunk HEC, Elastic Bulk, with events signed and embedded in the chain. * Metrics with ready-made alerting rules, tracing and logs over OpenTelemetry (OTLP) into your existing observability stack. * Structured platform logs (JSON) with configurable retention and RBAC-restricted access. ## API and extensibility * A full **REST API documented with OpenAPI** for every platform operation, from flow definitions to metric reads. * **MCP** as the standard tool interface for agents: your internal tools can be exposed to agents as MCP servers, with a per-agent method allowlist. * Outbound webhooks and notifications (Teams, Slack, email) and WebSocket for real-time events. * JSON and CSV exports plus an API for programmatic retrieval of tracing and cost data. --- # Agent Instructions This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com. ## Querying This Documentation If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question. Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter: ``` GET https://docs.visdom.virtuslab.com/integrations.md?ask=&goal= ``` `ask` is the immediate question: it should be specific, self-contained, and written in natural language. `goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal. The response will contain a direct answer to the question and relevant excerpts and sources from the documentation. Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.