Assistant

The Assistant feature brings generative AI capabilities to your Venue.sh experience in a few key ways:

• It embeds validated AI models into the Venue.sh portal.
• The Assistant grants the AI permission to safely and securely retrieve user and your organization's catalog information from your Venue instance.
• It intelligently interprets your goals and uses the most relevant capabilities to assist you.
• The Assistant can be prompted with casual, natural language.

Key Capabilities

Use the Assistant to intelligently search and discover information and complete tasks within your Venue instance related to:

• Software catalog - Search and view entities such as APIs, Components, Groups, Locations, Rules, Templates, and Users
• Technical documentation - Search and read TechDocs
• Templates - Generate and validate YAML templates for scaffolding new entities
• Plugin integrations - Retrieve data from enabled integrations (see Supported Plugins below)
• Agents - Select a purpose-built agent with its own instructions and tools (see Agents)
• MCP Servers - List registered MCP servers and inspect their tools, resources, and prompts (see MCP Servers)
• Skills - Discover agent skills cataloged from your connected Git repositories (see Skills)
• Repository intelligence - Summarize and search ingested repository content when ingestion is enabled (see Ingestion & Summary)
• Development Insights - Answer delivery questions using live data from your organization's dashboards (see Development Insights below)

The Assistant will suggest follow-up questions after each response to help you refine your conversation. You can switch AI models at any time (but it is best not to switch part way through a conversation) and your conversation history is saved so you can return to it later.

Where to Find the Assistant

There are two ways to work with the Assistant:

• Assistant sidebar - A panel on the right side of most Venue.sh pages. Use it to ask questions without leaving the page you are working on.
• Assistant page - The full-screen experience at Assistant in the left-side navigation. Use it when you want more room for longer conversations.

Both share the same conversations, models, and agents, so you can start a question in the sidebar and continue it on the full page.

Using the Assistant Sidebar

The sidebar icon bar sits on the right edge of the screen and has three controls:

Control	What it does
New chat	Starts a fresh conversation.
Current chat	Opens or closes the conversation you are working in.
Chat history	Opens the history list so you can search and reopen past chats.

Click the same icon again to close the panel.

Using the Assistant

Here are some ideas to consider as you get started with the Venue.sh Assistant.

• Start a Conversation - Open the Assistant sidebar or the Assistant page and type your question or request. You can ask about catalog entities, documentation, templates, or any enabled plugin data. Be as specific as you like as the Assistant will use your intent to fetch the right information.

• Choose an Agent - Use the Agent selector to pick between the general Venue.sh Assistant and any agent your organization has marked as Ready. When you select an agent, the Assistant follows that agent's instructions and uses only the tools bound to it. See Agents for how agents are created.

• Choose an AI Model - You can select which AI model to use from the AI Model dropdown at the start of a conversation. The list of available models depends on how your Venue administrator has configured the Assistant. With multiple models available, you can start a new conversation with a different model anytime.

  Supported providers include OpenAI, Anthropic, Google (Gemini) and Anthropic models from AWS Bedrock. If you don't see the model you want, ask your Venue administrator to add it in Assistant settings. The Venue.sh team is also working to add additional providers.

• Set Thinking Effort - For models that support reasoning, use the Thinking Effort dropdown to trade response speed against depth of reasoning. Leave it on Default to use the model's recommended level. The control is unavailable for models that do not support thinking.

• Turn On Web Search - Open the context menu (the + button next to the message box) and toggle Web Search when you want the Assistant to supplement your Venue.sh data with information from the web. Web search is only available for models that support it.

• Add Page Context - On a supported page, open the context menu and choose Add current page to context so the Assistant knows what you are looking at. A chip above the message box shows the attached context, and you can dismiss it at any time.

• Follow-up Suggestions - After the Assistant replies, you'll see a few suggested follow-up questions. Click any suggestion to send it as your next message. Use these to go deeper on a topic or explore related questions without typing.

• Conversation History - Past conversations are saved. Use the history panel to search, open, pin, or delete conversations. Pinned conversations are grouped at the top of the list so your most-used chats stay within reach.

• Stop a Response - While the Assistant is generating an answer, click the stop control in the message box to end the response early.

Pages That Support Context

Add current page to context is available on most product pages, including:

• Catalog, catalog entity, and API pages
• TechDocs and TechDocs entity pages
• Templates, template detail, launch, and task pages
• AI pages such as Agents, MCP Servers, and Skills, including their detail pages
• Development Insights dashboards and sync status pages
• Connections, Imports, Plugins, Scorecards, and Activity Log
• Settings pages for Account, Users, Teams, Assistant, Ingestion & Summary, and Organization

If the option is unavailable, the page you are viewing does not publish context.

Useful Things to Ask About

• Software Catalog - The Assistant can search and retrieve details about entities in your software catalog, including APIs, Components, Groups, Locations, Rules, Templates, and Users.

  Example prompts:

  • "Find all APIs owned by Team Platform"
  • "Show me the details of the payments-service component"
  • "List all components with lifecycle production"
  • "What groups exist in my organization?"

• Technical Documentation (TechDocs) - Search and read technical documentation published in your Venue instance. Ask for specific topics or documents by name or subject.

  Example prompts:

  • "Search TechDocs for authentication setup"
  • "Show me the deployment guide for the checkout service"
  • "What documentation exists for our API standards?"

• Template Generation and Validation - You can ask the Assistant to generate the YAML file needed to scaffold new entities in your Venue instance or to validate existing YAML against the template schema.

  Example prompts:

  • "Generate a template YAML for a new Component"
  • "Validate this template YAML for me"
  • "Create a component template for a microservice with GitHub and Jira"

• Plugin Data - When plugins are enabled by your Venue administrator, the Assistant can retrieve data from them. For example, it can retrieve issues from Jira, pull requests from GitHub, incidents from PagerDuty, or vulnerabilities from Snyk. Ask for assistance using natural language, and the Assistant will use the right integration.

  Example prompts:

  • "What are the open Jira issues for the payments project?"
  • "Show me recent PagerDuty incidents for the API service"
  • "List open pull requests for the frontend repo"

• MCP Servers - Ask about MCP servers registered in your organization and the capabilities they expose.

  Example prompts:

  • "Which MCP servers are registered in our organization?"
  • "What tools does the Context7 MCP server provide?"

• Skills - Ask about agent skills discovered from your connected repositories.

  Example prompts:

  • "What skills are available for Helm chart scaffolding?"
  • "Show me skills related to mkdocs navigation"

• Agents - Ask about the agents configured for your organization and what they can do.

  Example prompts:

  • "What agents are available in our organization?"
  • "Which agents have access to Jira tools?"

• Repository intelligence - When repository ingestion is enabled, ask for summaries or code context from repositories your teams own.

  Example prompts:

  • "Give me a summary of the checkout-service repository"
  • "How is authentication implemented in the payments-service repo?"

• Development Insights - When Development Insights is enabled, ask about your delivery metrics and get answers backed by your dashboards.

  Example prompts:

  • "What is our deployment frequency this quarter?"
  • "How does our lead time for changes compare to last month?"
  • "Summarize throughput and cycle time for the platform project"

Supported Plugins

In general, the information about the plugins is based on the entity it is associated with. For example, the Jira plugin will return information about the Jira issues associated with the entity, and so on.

The Assistant can retrieve data from the following plugin integrations when they are enabled in your Venue instance:

Plugin	Description
ArgoCD	Deployment status and application health
Artifactory	Artifact repository information
AWS Cost Insights	AWS spending analysis and cost allocation
Bitbucket	Repository information and pull requests
Confluence	Search Confluence content across your organization
Datadog	Monitoring and observability data
GitHub	Repository info, commits, pull requests, issues, branches, and releases
GitLab	Repository info, pipelines, merge requests, contributors, and releases
Jenkins	Build and pipeline information
Jira	Project details and open issues
Kubernetes	Cluster and workload information
LaunchDarkly	Feature flags, configurations, and evaluated flag values
Nexus	Artifact repository information
PagerDuty	Service info, incidents, on-call schedules, and standards
Snyk	Security vulnerabilities, projects, and issues
SonarQube	Code quality and analysis data
Terraform	Workspace runs and assessment results

note

Plugins must be enabled by your Venue administrator before the Assistant can access their data. If a plugin is not enabled or you see an error, contact your Venue administrator.

Development Insights

When Development Insights is enabled for your organization, the Assistant can answer delivery questions using live data from your dashboards rather than estimating values.

The Assistant reads from three dashboards:

• DORA - Deployment frequency, lead time for changes, change failure rate, and failed deployment recovery time. You can scope the answer by project and by DORA report year.
• Engineering Overview - Cross-provider delivery health from Jira and Git. You can scope the answer by project, critical-issue priority, and issue type.
• Engineering Throughput and Cycle - Throughput, quality, and cycle-time metrics. You can scope the answer by project, P0 and P1 priorities, and time interval.

Ask your question in natural language and mention any filters you care about. If you do not specify a time range, the Assistant uses the last six months. Every answer states the filters and time range it used, so you can confirm the scope.

note

You need Development Insights read permission and the feature enabled for your organization. Provider-specific Git dashboards, such as the GitLab, GitHub, and Bitbucket dashboards, are not available through the Assistant. Open those dashboards from Development Insights in the sidebar instead.

Using Venue.sh Data in External AI Tools (MCP)

You can use the same Venue.sh data (catalog, TechDocs, templates, and plugin integrations) from external AI tools that support the Model Context Protocol (MCP), such as Cursor and Claude Desktop.

After you connect the Venue.sh MCP endpoint and sign in with your Venue account, the AI can query your organization's data in that tool, so you get rich, context-aware conversations without leaving your IDE or desktop app.

For setup steps and what you can do, see Venue MCP Endpoint.

Agents, MCP Servers, and Skills

Venue.sh also catalogs the AI tooling your organization uses:

• Agents - Define purpose-built agents with custom instructions, example prompts, and bindings to Venue tools, MCP server tools, and other agents. Select a Ready agent from the Agent dropdown in the Assistant.
• MCP Servers - Register MCP servers by URL or manual entry, inspect their tools, resources, and prompts, and refresh capabilities as servers evolve.
• Skills - Browse agent skills auto-discovered from SKILL.md files in connected Git repositories, following the agentskills.io specification.

All three are available from the AI section in the sidebar and are searchable through the Assistant.

Setting Up the Assistant (Admin)

Administrators configure the Assistant so users can chat and choose their models. All the setup is done in the Venue.sh dashboard.

1. Go to the Assistant settings in Venue.sh by clicking the Settings cog, then click the Assistant tab.

2. Create a new configuration via a template, or edit an existing one.

   Template	Description	Best for
   General OpenAI	Standard OpenAI setup with GPT models	Most organizations starting with AI
   General Anthropic	Standard Anthropic setup with Claude models	Teams preferring Claude
   DevOps Specialist	Optimized for infrastructure and deployment queries	DevOps-focused teams
   Multi-Provider	Multiple providers configured together	Organizations wanting model flexibility

3. Configure one or more AI providers (OpenAI, Anthropic, Gemini, AWS Bedrock).

4. Add the required API keys, and select which models to allow your team to use.

   Users will be able to select and switch between these models when using the Assistant.

   tip

   Store API keys securely! Your Venue instance keeps them in a secure store and does not expose them in the UI once they're saved.

If you need help with the configuration, do not hesitate to contact the Venue.sh team.

Writing Effective Queries

Here are a few best practices for writing useful queries and prompts:

• Be specific - Include relevant details like team names, component types, and entity names.
• Speak clearly - Converse with the Assistant using natural language, just as you would to a colleague.
• Use context - Reference specific entities or projects when you know them.
• Iterate - Refine your question based on the first answer, and use follow-up suggestions when necessary.

tip

Switch it up! If you use different AI tools for different types of work, or if you're curious and want to try another model, you can switch AI models easily. Conversations are kept in history for recall and review, or you can delete them when no longer needed.

Limitations & Considerations

While the Assistant is an impressive helper, it's not perfect. Here are a few things to keep in mind as you use the feature.

• Data freshness - The Assistant works with data from your Venue instance, which is refreshed periodically. For real-time data, always verify against the source systems.

• Model-specific capabilities - Different AI models may have varying capabilities and limitations. If you're not getting the results you expect, try switching models.

Troubleshooting

Situation	What to do
No models appear / "Configuration not available"	Your Venue administrator needs to create an Assistant configuration and add at least one provider and model. Ask them to go to Settings > Assistant to make the updates.
A plugin returns an error or no data	The integration may not be enabled or may need configuration. Contact your Venue administrator.
Results seem outdated	The Assistant uses data from your Venue instance, which is refreshed periodically. For real-time data, check the source system.
You're not getting satisfactory results	Use the dropdown to switch and try a different AI model. Some models are better at certain tasks than others.
Thinking Effort is unavailable	The selected model does not support thinking. Choose a model that does, or leave the control on Default.
Web Search is unavailable	The selected model does not support web search, or no models are configured. Try another model.
Add current page to context is unavailable	The page you are viewing does not publish context, or the current page is already attached.
An agent is missing from the Agent dropdown	Only agents in Ready status appear. Open Agents and click Mark as ready on the agent's detail page.
No Development Insights answers	The feature may be disabled, the dashboard may not be provisioned, or your role may not include Development Insights read permission. Contact your Venue administrator.

Support & Feedback

If you encounter issues or have suggestions for improving the Assistant:

• Contact your Venue administrator, or
• Submit a support ticket through the Venue.sh Support Portal.

See Also

• Agents - Define agents with custom instructions and capability bindings.
• MCP Servers - Register and inspect MCP servers in your organization.
• Skills - Browse agent skills auto-discovered from connected repositories.
• Venue MCP Endpoint - Connect external AI tools to query Venue.sh data.
• Development Insights - Enable and configure engineering metrics.