> ## Documentation Index > Fetch the complete documentation index at: https://docs.mindset.ai/llms.txt > Use this file to discover all available pages before exploring further. # Create a new agent > Create a new agent with full configuration options including personality, formatting, policies, and memory settings. ## OpenAPI ````yaml /deploy/api/openapi.json post /api/v2/appuid/{appUid}/agents openapi: 3.0.1 info: title: Mindset API Gateway version: 1.2.0 servers: - url: https://a.api.mindset.ai - url: https://b.api.mindset.ai - url: https://c.api.mindset.ai - url: https://d.api.mindset.ai security: [] tags: - name: Agent Configuration - name: Agent Sessions - name: AppStatus - name: Context Files - name: Contexts - name: DynamicComponents - name: Labels - name: SDK Users - name: Threads - name: Webhooks - name: Agent Management (deprecated) paths: /api/v2/appuid/{appUid}/agents: post: tags: - Agent Configuration summary: Create a new agent description: >- Create a new agent with full configuration options including personality, formatting, policies, and memory settings. operationId: createAgentV2 parameters: - name: appUid in: path description: Your application's uid required: true schema: minLength: 1 type: string requestBody: content: application/json: schema: required: - agentName - agentPurpose type: object properties: agentName: maxLength: 50 minLength: 1 type: string description: Name of the agent agentPurpose: maxLength: 1000 minLength: 1 type: string description: Purpose of the agent templateAgentUid: type: string description: >- UID of an existing agent to use as a template for sources presentation options (citations, references etc) personality: $ref: '#/components/schemas/PersonalityObject' formatting: $ref: '#/components/schemas/FormattingObject' policies: $ref: '#/components/schemas/PoliciesObject' additionalPolicyRules: maxLength: 500 type: string description: Additional free-form policy rules additionalInformation: maxLength: 65535 type: string description: Additional context information for the agent iceBreakerMessage: maxLength: 1000 type: string description: >- Invisible instruction to prompt the agent to greet users in a specific way memory: $ref: '#/components/schemas/MemoryObject' options: $ref: '#/components/schemas/SDK3OptionsObject' live: type: boolean description: Agent is active and live for end users default: false showPromptLibrary: type: boolean description: Show prompt library in SDK UI default: false contextUids: maxItems: 30 type: array description: Array of existing context UIDs (max 30) items: type: string mcpserverUids: maxItems: 5 type: array description: Array of existing MCP server UIDs (max 5) items: type: string planUids: maxItems: 20 type: array description: Array of existing plan UIDs (max 20) items: type: string skillUids: maxItems: 20 type: array description: Array of existing skill UIDs (max 20) items: type: string tags: maxItems: 10 type: array description: Array of tags (max 10) items: maxLength: 60 minLength: 1 type: string access: type: string description: Access control mode for this agent default: restrictedAccess enum: - openAccess - restrictedAccess externalTenantId: maxLength: 100 type: string description: >- External tenant identifier (alphanumeric, underscore, hyphen only) required: true responses: '201': description: Agent created successfully content: application/json: schema: $ref: '#/components/schemas/AgentResponse' '400': description: Bad Request - Invalid input content: {} '401': description: Unauthorized content: {} '403': description: Forbidden content: {} '404': description: Not Found - Template agent does not exist content: {} '422': description: Unprocessable Entity - Validation error content: {} '500': description: Internal Server Error content: {} security: - APIKeyHeader: [] components: schemas: PersonalityObject: type: object properties: preset: type: string description: >- Personality preset. Use 'custom' to provide your own personality text. default: friendly enum: - friendly - professional - analytical - strict - custom customText: maxLength: 500 minLength: 1 type: string description: >- Custom personality text (required if preset is 'custom', ignored otherwise) description: Agent personality configuration FormattingObject: type: object properties: preset: type: string description: >- Formatting preset. Use 'custom' to provide your own formatting instructions. default: conversational enum: - conversational - structured - storytelling - technical - custom customText: maxLength: 500 minLength: 1 type: string description: >- Custom formatting text (required if preset is 'custom', ignored otherwise) description: Agent response formatting configuration PoliciesObject: type: object properties: focusAndScope: $ref: '#/components/schemas/FocusAndScopePolicy' knowledgeLimits: $ref: '#/components/schemas/KnowledgeLimitsPolicy' nonContextSegmentMessage: $ref: '#/components/schemas/NonContextSegmentMessagePolicy' clarification: $ref: '#/components/schemas/ClarificationPolicy' recommendations: $ref: '#/components/schemas/RecommendationsPolicy' neutrality: $ref: '#/components/schemas/NeutralityPolicy' professionalConduct: $ref: '#/components/schemas/ProfessionalConductPolicy' legalFinancial: $ref: '#/components/schemas/LegalFinancialPolicy' safeguarding: $ref: '#/components/schemas/SafeguardingPolicy' feedback: $ref: '#/components/schemas/FeedbackPolicy' description: Agent policy configuration with 10 toggleable policies MemoryObject: type: object properties: allowAccess: type: boolean description: Allow agent to access memory default: false allowContribution: type: boolean description: Allow agent to contribute to memory default: false miningGuidance: maxLength: 500 type: string description: Guidance for memory mining description: >- Agent memory configuration. Note: Memory access must be enabled at the app level for these settings to take effect. SDK3OptionsObject: type: object properties: showSourceContent: type: boolean description: >- Show segments and files from knowledge contexts that informed the response default: false sourceDisplayStyle: type: string description: >- How sources are displayed. 'compact' = expandable list, 'card' = carousel default: compact enum: - compact - card sourceCardType: type: string description: >- What to show as cards in carousel. Only present when sourceDisplayStyle is 'card' default: segments enum: - segments - files showSpeakers: type: boolean description: Show speaker names from audio and video files default: false showFollowUpQuestions: type: boolean description: Show AI-generated follow-up questions based on source topics default: false showCitations: type: boolean description: Show citations within response linking to sources default: false showQuickReplies: type: boolean description: Display clickable buttons when agent suggests options default: true description: >- SDK3 display options. Only present for apps with SDK3 enabled (labsSDKV3Beta: true). AgentResponse: type: object properties: uid: type: string description: Unique ID of the agent agentName: type: string description: Name of the agent agentPurpose: type: string description: Purpose of the agent personality: $ref: '#/components/schemas/PersonalityObject' formatting: $ref: '#/components/schemas/FormattingObject' policies: $ref: '#/components/schemas/PoliciesObject' additionalPolicyRules: type: string description: Additional free-form policy rules additionalInformation: type: string description: Additional context information for the agent iceBreakerMessage: type: string description: >- Invisible instruction to prompt the agent to greet users in a specific way memory: $ref: '#/components/schemas/MemoryObject' options: $ref: '#/components/schemas/SDK3OptionsObject' live: type: boolean description: Agent is active and live for end users showPromptLibrary: type: boolean description: Show prompt library in SDK UI contextUids: type: array description: Array of context UIDs items: type: string mcpserverUids: type: array description: Array of MCP server UIDs items: type: string planUids: maxItems: 20 type: array description: Array of plan UIDs (max 20) items: type: string skillUids: maxItems: 20 type: array description: Array of skill UIDs (max 20) items: type: string tags: type: array description: Array of tags items: type: string access: type: string description: Access control mode enum: - openAccess - restrictedAccess externalTenantId: type: string description: External tenant identifier createdAt: type: string description: Creation timestamp (ISO 8601) format: date-time FocusAndScopePolicy: type: object properties: enabled: type: boolean default: false scope: maxLength: 65535 type: string description: >- The scope of topics the agent should focus on (required when enabled) description: Focus and Scope policy - restricts agent to specific topics KnowledgeLimitsPolicy: type: object properties: enabled: type: boolean default: false themeSpecificAllowance: maxLength: 65535 type: string description: Themes the agent can discuss when out of context (optional) description: Knowledge Limits and Apologies policy - handles out-of-context queries NonContextSegmentMessagePolicy: type: object properties: enabled: type: boolean default: false message: maxLength: 65535 type: string description: Message to display for non-context segments (required when enabled) description: >- Non-Context Segment Message policy - custom message for out-of-context responses ClarificationPolicy: type: object properties: enabled: type: boolean default: false description: >- Clarification and User Understanding policy - helps users understand responses better RecommendationsPolicy: type: object properties: enabled: type: boolean default: false description: >- Recommendations and Examples policy - provides suggestions and examples in responses NeutralityPolicy: type: object properties: enabled: type: boolean default: false description: Neutrality policy - uses gender-neutral speaker references ProfessionalConductPolicy: type: object properties: enabled: type: boolean default: false description: >- Professional Conduct and Bias policy - maintains professional tone and avoids bias LegalFinancialPolicy: type: object properties: enabled: type: boolean default: false description: >- Avoid Legal and Financial Advice policy - prevents giving legal or financial guidance SafeguardingPolicy: type: object properties: enabled: type: boolean default: false email: maxLength: 65535 type: string description: Contact email for safeguarding issues (required when enabled) format: email description: Safeguarding policy - handles sensitive/mental health disclosures FeedbackPolicy: type: object properties: enabled: type: boolean default: false description: Feedback Encouragement policy - encourages users to provide feedback securitySchemes: APIKeyHeader: type: apiKey name: x-api-key in: header ````