{"passport":{"unfragile":{"@version":"1.0","version":"2026-05","artifact":{"id":"create-llama","slug":"create-llama","name":"create-llama","type":"cli","url":"https://github.com/run-llama/create-llama","page_url":"https://unfragile.ai/create-llama","categories":["app-builders","rag-knowledge","documentation"],"tags":[],"pricing":{"model":"free","free":true,"starting_price":null},"status":"active","verified":false},"capabilities":[{"id":"create-llama__cap_0","uri":"capability://automation.workflow.interactive.cli.guided.project.scaffolding","name":"interactive-cli-guided-project-scaffolding","description":"Provides a command-line interface that walks developers through a series of prompts to configure and generate a complete LlamaIndex application. The CLI uses a template system that reads user selections (framework choice, LLM provider, vector database, use case) and dynamically renders the appropriate boilerplate code by composing pre-built template fragments. Supports both quick-start mode with sensible defaults and pro mode for granular component selection.","intents":["I want to bootstrap a new LLM application without writing boilerplate configuration","I need to quickly choose between Next.js, FastAPI, Express, or LlamaIndex Server for my RAG project","I want to configure my LLM provider, vector database, and document types through guided prompts rather than manual setup"],"best_for":["developers new to LlamaIndex who want zero-config startup","teams prototyping RAG applications and need to iterate on architecture quickly","solo developers building proof-of-concept LLM agents without DevOps overhead"],"limitations":["CLI-driven setup means limited programmatic control — no SDK for headless generation","Template composition is static at generation time — runtime template switching not supported","Pro mode requires understanding of all available components; no intelligent recommendations based on use case"],"requires":["Node.js 16+ or Python 3.8+ depending on target framework","npm, yarn, pnpm, or bun package manager","API keys for selected LLM provider (OpenAI, Anthropic, etc.) and vector database"],"input_types":["CLI prompts (text selections)","environment variables for API keys"],"output_types":["complete project directory structure","configured package.json with dependencies","environment template files (.env.example)","runnable application code"],"categories":["automation-workflow","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_1","uri":"capability://code.generation.editing.multi.framework.application.generation","name":"multi-framework-application-generation","description":"Generates production-ready applications across four distinct backend frameworks (Next.js full-stack, FastAPI Python backend, Express Node.js backend, LlamaIndexServer) from a unified template abstraction. Each framework template includes pre-configured routing, middleware, streaming endpoints, and document upload handlers specific to that framework's patterns. The generation process selects the appropriate template tree based on user choice and renders it with injected configuration.","intents":["I want to generate a Next.js full-stack app with LlamaIndex integrated","I need a Python FastAPI backend with a separate Next.js frontend for my RAG pipeline","I want to use Express.js with LlamaIndex.TS instead of Python"],"best_for":["teams with existing Next.js infrastructure who want to add RAG capabilities","Python-first teams who prefer FastAPI for backend services","Node.js teams avoiding Python dependencies","teams deploying to LlamaIndex Server for managed inference"],"limitations":["Each framework template must be maintained separately — adding a feature requires updates across 4 codebases","Framework-specific patterns mean generated code is not easily portable between frameworks","LlamaIndexServer template is specialized and may have fewer customization options than open-source alternatives"],"requires":["Node.js 16+ for Next.js and Express templates","Python 3.8+ for FastAPI template","npm/yarn/pnpm for JavaScript templates, pip for Python template"],"input_types":["framework selection (nextjs | fastapi | express | llamaindexserver)","configuration options (LLM provider, vector DB, use case)"],"output_types":["Next.js: app directory structure with API routes, components, and pages","FastAPI: main.py with routers, dependencies, and async endpoints","Express: server.ts with route handlers and middleware","LlamaIndexServer: server configuration with workflow definitions"],"categories":["code-generation-editing","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_10","uri":"capability://automation.workflow.project.dependency.management.and.lockfile.generation","name":"project-dependency-management-and-lockfile-generation","description":"Generates package.json (or requirements.txt for Python) with all required dependencies for the selected framework, LLM providers, vector databases, and tools, pinned to compatible versions. Includes development dependencies for testing, linting, and build tools. Generates lockfiles (pnpm-lock.yaml, package-lock.json, poetry.lock) ensuring reproducible builds across environments. Handles dependency resolution for complex transitive dependencies.","intents":["I want all dependencies automatically included without manually researching versions","I need reproducible builds with locked dependency versions","I want to avoid dependency conflicts between LlamaIndex, vector databases, and other packages"],"best_for":["teams deploying to production and needing reproducible builds","developers avoiding dependency hell and version conflicts","applications with complex dependency graphs (multiple vector databases, tools, etc.)"],"limitations":["Dependency versions are fixed at generation time — updating requires manual package.json edits or regeneration","No automatic dependency upgrade checking — security updates require manual intervention","Lockfile generation is framework-specific (npm, yarn, pnpm, poetry) with different formats","Transitive dependency conflicts may still occur with incompatible versions of underlying libraries"],"requires":["npm, yarn, pnpm, or bun for JavaScript projects","pip or poetry for Python projects"],"input_types":["framework selection (determines base dependencies)","service selections (LLM provider, vector database, tools)"],"output_types":["package.json with all dependencies and versions","lockfile (pnpm-lock.yaml, package-lock.json, poetry.lock, etc.)","optional: requirements.txt for Python"],"categories":["automation-workflow","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_11","uri":"capability://code.generation.editing.typescript.python.type.safety.generation","name":"typescript-python-type-safety-generation","description":"Generates TypeScript type definitions and Python type hints for all API contracts, data models, and function signatures. For TypeScript projects, generates strict tsconfig.json with strict mode enabled. For Python projects, generates Pydantic models for request/response validation. Includes type definitions for chat messages, document metadata, and tool parameters matching the backend API schema.","intents":["I want type safety across my frontend and backend without manual type definition","I need Pydantic models for request validation in my FastAPI application","I want TypeScript strict mode enabled to catch type errors at compile time"],"best_for":["teams prioritizing type safety and compile-time error detection","applications with complex data models requiring validation","developers wanting IDE autocomplete and type checking across API boundaries"],"limitations":["Type definitions are generated from templates — complex types require manual refinement","TypeScript strict mode may require additional type annotations in generated code","Pydantic models are generated from API schema — runtime validation errors still possible","Type definitions are static at generation time — schema changes require regeneration or manual updates"],"requires":["TypeScript 4.5+ for TypeScript projects","Python 3.8+ with Pydantic for Python projects"],"input_types":["API schema (chat messages, documents, tools, etc.)","data model definitions"],"output_types":["TypeScript type definitions (.d.ts files)","Python Pydantic models","tsconfig.json with strict mode","type-safe API client code"],"categories":["code-generation-editing","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_12","uri":"capability://automation.workflow.ci.cd.workflow.and.deployment.configuration","name":"ci-cd-workflow-and-deployment-configuration","description":"Generates GitHub Actions workflows (or equivalent CI/CD configuration) for testing, building, and deploying the generated application. Includes workflows for running tests, linting, type checking, building Docker images, and deploying to cloud platforms (Vercel for Next.js, cloud run for FastAPI, etc.). Supports environment-specific deployments with secret management integration.","intents":["I want automated testing and linting on every pull request","I need CI/CD pipelines to deploy my application to production","I want Docker images built automatically for my FastAPI backend"],"best_for":["teams deploying to production and needing automated CI/CD","applications requiring testing and linting before deployment","teams using GitHub for version control and wanting native GitHub Actions"],"limitations":["CI/CD workflows are GitHub Actions specific — other CI/CD systems require manual conversion","Deployment targets are limited to common platforms (Vercel, Cloud Run, etc.) — custom deployment requires modification","Secret management integration requires manual setup in GitHub repository settings","Workflow optimization for build time and resource usage requires manual tuning"],"requires":["GitHub repository with Actions enabled","deployment platform credentials (Vercel token, Google Cloud credentials, etc.)","Docker for containerized deployments"],"input_types":["framework selection (determines build and test commands)","deployment target (vercel | cloud-run | etc.)"],"output_types":[".github/workflows/test.yml",".github/workflows/deploy.yml","Dockerfile for containerized deployments","deployment configuration files"],"categories":["automation-workflow","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_2","uri":"capability://memory.knowledge.vector.database.integration.configuration","name":"vector-database-integration-configuration","description":"Generates application code with pre-configured vector database clients and connection logic for multiple vector store backends (MongoDB, PostgreSQL, Pinecone, Weaviate, Milvus, etc.). The generation process injects database-specific initialization code, embedding model configuration, and index creation logic into the generated application. Supports both local development databases and cloud-hosted services with environment-based credential injection.","intents":["I want to use Pinecone for vector storage in my RAG application without manual client setup","I need to switch from Pinecone to a self-hosted PostgreSQL vector database","I want my generated app to automatically create and manage vector indices on startup"],"best_for":["teams evaluating different vector databases and want quick integration testing","developers building RAG systems who want database selection decoupled from application code","teams with existing vector database infrastructure wanting to integrate with LlamaIndex"],"limitations":["Vector database selection is made at generation time — switching databases requires regenerating the project","No built-in migration tooling for moving embeddings between vector stores","Advanced vector database features (hybrid search, metadata filtering) require manual code additions after generation","Embedding model is fixed at generation time; changing embedding providers requires code modification"],"requires":["API credentials for selected vector database (Pinecone API key, PostgreSQL connection string, etc.)","Vector database service running or accessible (local or cloud)","LlamaIndex Python or TypeScript SDK with vector store integrations"],"input_types":["vector database selection (pinecone | mongodb | postgresql | weaviate | milvus | etc.)","database credentials (API keys, connection strings, endpoints)"],"output_types":["initialized vector store client code","index creation and initialization logic","environment variable templates for database credentials","embedding model configuration"],"categories":["memory-knowledge","data-processing-analysis"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_3","uri":"capability://data.processing.analysis.document.ingestion.pipeline.generation","name":"document-ingestion-pipeline-generation","description":"Generates a document upload and processing pipeline that accepts multiple file formats (PDF, text, CSV, Markdown, Word, HTML, and for Python: video and audio) and automatically indexes them into the vector database. The generated code includes file type detection, document parsing using LlamaIndex document loaders, chunking strategy configuration, and embedding generation. Provides both API endpoints for programmatic upload and UI components for user-facing document management.","intents":["I want users to upload documents through a web UI and have them automatically indexed for RAG","I need to ingest multiple document formats (PDFs, CSVs, Markdown) into my knowledge base","I want to support video and audio file ingestion in my Python application"],"best_for":["teams building document-based RAG systems with user-facing upload interfaces","applications requiring multi-format document support (mixed PDF, text, video)","developers who want chunking and embedding strategies pre-configured"],"limitations":["Chunking strategy is fixed at generation time — changing chunk size or overlap requires code modification","No built-in document deduplication or update detection — re-uploading same document creates duplicates","Video/audio processing (Python only) requires additional dependencies and may have latency implications","No streaming ingestion for large document batches — all processing happens synchronously"],"requires":["LlamaIndex document loaders for target file types","Vector database configured and accessible","For video/audio: ffmpeg or equivalent media processing library (Python only)","Sufficient storage for document cache and embeddings"],"input_types":["uploaded files (PDF, TXT, CSV, MD, DOCX, HTML, MP4, MP3, etc.)","file metadata (filename, upload timestamp)"],"output_types":["parsed document nodes with metadata","generated embeddings stored in vector database","document index with retrieval metadata","API response confirming ingestion status"],"categories":["data-processing-analysis","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_4","uri":"capability://text.generation.language.streaming.chat.endpoint.generation","name":"streaming-chat-endpoint-generation","description":"Generates a streaming chat API endpoint that accepts conversation history and user messages, processes them through the LlamaIndex RAG pipeline, and returns responses as server-sent events (SSE) or streaming JSON. The generated endpoint includes context window management, prompt templating, and streaming response handling specific to the chosen LLM provider. Supports both stateless request-response and stateful conversation management with optional persistence.","intents":["I want a chat endpoint that streams responses from my RAG system in real-time","I need to maintain conversation history across multiple chat turns","I want to use different LLM providers (OpenAI, Anthropic, local models) with the same chat interface"],"best_for":["teams building conversational RAG interfaces with real-time response requirements","applications needing multi-turn conversation with context awareness","developers wanting to experiment with different LLM providers without code changes"],"limitations":["Conversation history is passed in each request — no server-side session persistence by default","Context window management is basic — no automatic summarization for long conversations","Streaming implementation is framework-specific (Next.js streaming vs FastAPI SSE vs Express) with different latency characteristics","No built-in rate limiting or conversation quota management"],"requires":["LLM provider API key (OpenAI, Anthropic, etc.) or local model endpoint","Vector database with indexed documents","Framework-specific streaming support (Next.js Response, FastAPI StreamingResponse, Express res.write)"],"input_types":["conversation history (array of messages with role and content)","user message (text)","optional: system prompt, temperature, max_tokens"],"output_types":["streaming response (SSE or chunked JSON)","complete message text","optional: token usage metadata, source documents"],"categories":["text-generation-language","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_5","uri":"capability://tool.use.integration.llm.provider.abstraction.and.configuration","name":"llm-provider-abstraction-and-configuration","description":"Generates application code with pluggable LLM provider configuration that supports OpenAI, Anthropic, local models, and other LlamaIndex-supported providers. The generated code uses LlamaIndex's LLM abstraction layer to decouple provider-specific logic from application code, allowing provider switching via environment variables without code changes. Includes provider-specific configuration (temperature, max_tokens, model selection) injected at initialization.","intents":["I want to use OpenAI by default but easily switch to Anthropic or a local model","I need to configure different LLM parameters (temperature, max_tokens) per environment","I want to support multiple LLM providers and let users choose at runtime"],"best_for":["teams evaluating different LLM providers and wanting to switch without code changes","applications needing cost optimization by routing to cheaper models for certain tasks","developers building multi-tenant systems where different users use different LLM providers"],"limitations":["Provider-specific features (function calling, vision, tool use) require conditional code paths","Cost tracking and usage monitoring must be implemented separately per provider","Fallback logic for provider outages is not generated — requires manual implementation","Model-specific prompt engineering may be needed for consistent quality across providers"],"requires":["API keys for selected LLM providers (OpenAI, Anthropic, etc.)","LlamaIndex SDK with provider integrations","Environment variable configuration for provider selection and credentials"],"input_types":["LLM provider selection (openai | anthropic | ollama | etc.)","model name (gpt-4, claude-3, llama2, etc.)","provider credentials (API keys, endpoints)"],"output_types":["initialized LLM client using LlamaIndex abstraction","environment variable templates for provider configuration","provider-specific configuration (temperature, max_tokens, etc.)"],"categories":["tool-use-integration","automation-workflow"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_6","uri":"capability://code.generation.editing.use.case.specific.template.selection","name":"use-case-specific-template-selection","description":"Provides pre-configured template variants for common LLM application patterns (RAG, agents, data analysis, report generation, etc.) that include domain-specific prompts, tool integrations, and workflow configurations. Each use case template includes example system prompts, relevant tool definitions, and architectural patterns optimized for that use case. Users select a use case during setup, and the generated application includes pre-wired components for that pattern.","intents":["I want a pre-configured RAG template that already has document retrieval wired up","I need an agent template with web search and code interpreter tools pre-integrated","I want a report generation template with structured output and formatting"],"best_for":["developers new to LLM applications who want working examples of common patterns","teams prototyping multiple use cases and wanting to start from relevant templates","non-technical founders building MVPs who want sensible defaults for their use case"],"limitations":["Use case templates are opinionated — customizing beyond the template requires code changes","Adding new use case templates requires maintaining additional template variants","Domain-specific prompts in templates may need tuning for specific data or requirements","Tool integrations in agent templates are basic examples — production use requires enhancement"],"requires":["understanding of the selected use case pattern","API keys for any tools included in the template (web search, code execution, etc.)","LlamaIndex SDK with tool/agent support"],"input_types":["use case selection (rag | agent | data-analysis | report-generation | etc.)"],"output_types":["template-specific application structure","pre-configured prompts and system messages","tool definitions and integrations","example workflow configurations"],"categories":["code-generation-editing","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_7","uri":"capability://tool.use.integration.agent.and.tool.integration.scaffolding","name":"agent-and-tool-integration-scaffolding","description":"Generates application code with pre-wired agent frameworks and tool integrations including web search, code interpreter, OpenAPI connectors, and custom tool definitions. The generated code includes agent initialization, tool registry setup, and function calling configuration specific to the chosen LLM provider. Supports both simple tool calling and complex multi-agent workflows with tool composition and error handling.","intents":["I want to create an agent that can search the web and answer questions","I need an agent with access to a code interpreter for data analysis","I want to connect my agent to external APIs through OpenAPI specifications"],"best_for":["teams building autonomous agents with external tool access","applications requiring multi-step reasoning with tool use","developers integrating LLM agents with existing API ecosystems"],"limitations":["Tool definitions are generated from templates — complex tool behavior requires manual implementation","Error handling for tool execution is basic — production use needs robust retry and fallback logic","Multi-agent workflows are not automatically orchestrated — requires manual coordination code","Tool calling support varies by LLM provider — some providers have limited function calling capabilities"],"requires":["LLM provider with function calling support (OpenAI, Anthropic, etc.)","API credentials for external tools (web search API, code execution service, etc.)","LlamaIndex agent and tool frameworks"],"input_types":["tool selection (web-search | code-interpreter | openapi-connector | custom)","tool configuration (API keys, endpoints, schemas)"],"output_types":["agent initialization code","tool registry with function definitions","function calling configuration for LLM provider","tool execution and error handling logic"],"categories":["tool-use-integration","planning-reasoning"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_8","uri":"capability://automation.workflow.environment.configuration.template.generation","name":"environment-configuration-template-generation","description":"Generates environment variable templates (.env.example) and configuration files with placeholders for all required credentials, API keys, and service endpoints. The generated templates include documentation for each variable explaining what it is and where to obtain it. Supports environment-specific configurations (development, staging, production) with different defaults and validation rules.","intents":["I want a template showing all environment variables my generated app needs","I need to configure different LLM providers and databases for dev vs production","I want documentation for each environment variable so my team knows what to set"],"best_for":["teams deploying generated applications to multiple environments","developers onboarding new team members who need to understand configuration","applications with complex credential management across multiple services"],"limitations":["Environment variable validation is not generated — requires manual implementation","No built-in secret management integration (AWS Secrets Manager, HashiCorp Vault, etc.)","Configuration is static at generation time — runtime configuration changes require code modification","No automatic credential rotation or expiration handling"],"requires":["understanding of required services and their credentials","environment-specific configuration values"],"input_types":["service selections (LLM provider, vector database, etc.)","environment targets (development, staging, production)"],"output_types":[".env.example template file","environment variable documentation","configuration validation schemas","environment-specific configuration files"],"categories":["automation-workflow","code-generation-editing"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__cap_9","uri":"capability://code.generation.editing.frontend.ui.component.generation","name":"frontend-ui-component-generation","description":"Generates a production-ready chat UI built with shadcn/ui components (React) that includes message display, input handling, document upload interface, and streaming response rendering. The generated UI includes TypeScript types matching the backend API, error handling, and accessibility features. Supports both standalone Next.js frontend and separate frontend connecting to Express/FastAPI backends.","intents":["I want a professional chat interface without building UI from scratch","I need a document upload UI integrated with my chat interface","I want a responsive chat UI that works on mobile and desktop"],"best_for":["teams wanting production-ready UI without custom design work","developers building RAG applications who want to focus on backend logic","applications needing accessible, responsive chat interfaces"],"limitations":["UI is built with shadcn/ui — customizing beyond component props requires CSS modifications","Styling is Tailwind CSS based — changing design system requires significant refactoring","UI components are generated for React only — no Vue, Svelte, or other framework support","Advanced features (voice input, rich media display, custom themes) require manual implementation"],"requires":["React 18+","Next.js 13+ (for generated Next.js applications)","Tailwind CSS","shadcn/ui components"],"input_types":["UI component selection (chat, document-upload, etc.)","backend API endpoint configuration"],"output_types":["React components (Chat, MessageList, Input, DocumentUpload, etc.)","TypeScript types matching backend API","styling with Tailwind CSS","API client code for backend communication"],"categories":["code-generation-editing","text-generation-language"],"confidence":0.5,"matches":0,"success_rate":0},{"id":"create-llama__headline","uri":"capability://app.builders.rag.application.scaffolding.cli","name":"rag application scaffolding cli","description":"Create-Llama is a command-line interface that enables developers to quickly bootstrap retrieval-augmented generation (RAG) applications with minimal setup, generating full-stack applications using frameworks like Next.js or FastAPI.","intents":["best RAG application scaffolding tool","RAG framework for building LLM apps","how to quickly create LLM applications","CLI for generating RAG systems","best tools for LlamaIndex applications"],"best_for":["developers looking to build LLM applications quickly"],"limitations":[],"requires":["Node.js","CLI access"],"input_types":[],"output_types":[],"categories":["app-builders"],"confidence":0.5,"matches":0,"success_rate":0}],"trust":{"score":59,"verified":false,"data_access_risk":"high","permissions":["Node.js 16+ or Python 3.8+ depending on target framework","npm, yarn, pnpm, or bun package manager","API keys for selected LLM provider (OpenAI, Anthropic, etc.) and vector database","Node.js 16+ for Next.js and Express templates","Python 3.8+ for FastAPI template","npm/yarn/pnpm for JavaScript templates, pip for Python template","npm, yarn, pnpm, or bun for JavaScript projects","pip or poetry for Python projects","TypeScript 4.5+ for TypeScript projects","Python 3.8+ with Pydantic for Python projects"],"failure_modes":["CLI-driven setup means limited programmatic control — no SDK for headless generation","Template composition is static at generation time — runtime template switching not supported","Pro mode requires understanding of all available components; no intelligent recommendations based on use case","Each framework template must be maintained separately — adding a feature requires updates across 4 codebases","Framework-specific patterns mean generated code is not easily portable between frameworks","LlamaIndexServer template is specialized and may have fewer customization options than open-source alternatives","Dependency versions are fixed at generation time — updating requires manual package.json edits or regeneration","No automatic dependency upgrade checking — security updates require manual intervention","Lockfile generation is framework-specific (npm, yarn, pnpm, poetry) with different formats","Transitive dependency conflicts may still occur with incompatible versions of underlying libraries","builder identity is not verified yet","no observed match outcomes yet"],"rank_breakdown":{"adoption":0.7,"quality":0.9,"ecosystem":0.6,"match_graph":0.25,"freshness":0.52,"weights":{"adoption":0.25,"quality":0.25,"ecosystem":0.1,"match_graph":0.28,"freshness":0.12}},"observed_outcomes":{"matches":0,"success_rate":0,"avg_confidence":0,"top_intents":[],"last_matched_at":null},"maintenance":{"status":"active","updated_at":"2026-06-17T09:51:04.690Z","last_scraped_at":null,"last_commit":null},"community":{"stars":null,"forks":null,"weekly_downloads":null,"model_downloads":null,"model_likes":null}},"distribution":{"claim_url":"https://unfragile.ai/submit?claim=create-llama","compare_url":"https://unfragile.ai/compare?artifact=create-llama"}},"signature":"JpSvU+ewAXCCW70jFI+S68v9u/DzlV+vwR3l6Foihy6lhxN/wWny6MMXTymVSpV/oabnfPr2ByQhS8V7968xCg==","signedAt":"2026-06-21T18:16:12.601Z","signedBy":"unfragile.ai","version":1},"_links":{"self":"https://unfragile.ai/api/v1/passport/create-llama","artifact":"https://unfragile.ai/create-llama","verify":"https://unfragile.ai/api/v1/verify?slug=create-llama","publicKey":"https://unfragile.ai/api/v1/trust-passport-public-key","spec":"https://unfragile.ai/trust","schema":"https://unfragile.ai/schema.json","docs":"https://unfragile.ai/docs"}}