AINativeainative.axAXΒ AUDIT
← All posts
AINative SDK Suite: Multi-Language Power for ZeroDB + Agent APIs
PLATFORM UPDATESDecember 24, 2025Β· 11 min read

AINative SDK Suite: Multi-Language Power for ZeroDB + Agent APIs

By Toby Morning
# AINative SDK Suite: Multi-Language Power for ZeroDB + Agent APIs We're thrilled to announce major updates to the **AINative SDK ecosystem**, delivering production-ready developer tools across **Python, TypeScript/JavaScript, and Go**. Whether you're building AI agents, semantic search systems, or data-intensive applications, we now have you covered with enterprise-grade SDKs in your language of choice. ## What's New Across All SDKs ### πŸš€ **Python SDK v1.1.0** - Now on PyPI **Installation:** `pip install zerodb-mcp` The Python SDK has reached production maturity with **60+ ZeroDB operations** and comprehensive type safety using Pydantic models. This release introduces groundbreaking **FREE embedding generation** via our ZeroDB Embeddings APIs. #### πŸ†• New Embeddings Operations (6 operations) ```python from zerodb_mcp import ZeroDBClient client = ZeroDBClient(api_key="your_api_key") # Generate embeddings using BAAI BGE models embedding = await client.embeddings.generate( text="Your document text here", model="BAAI/bge-small-en-v1.5" ) # One-step embedding + storage result = await client.embeddings.embed_and_store( project_id="your-project-id", text="Customer feedback: Great product!", metadata={"source": "feedback", "rating": 5} ) # Text-based semantic search (no manual embedding needed!) results = await client.embeddings.semantic_search( project_id="your-project-id", query="positive customer reviews", limit=10 ) ``` **Why This Matters:** Previously, developers had to pay for OpenAI or Cohere embeddings. Now you can generate **unlimited embeddings at zero cost** using state-of-the-art BAAI BGE models via our ZeroDB Embeddings APIs. #### Full ZeroDB Feature Set | Category | Operations | Capabilities | |----------|:----------:|--------------| | **Vector Operations** | 10 | Upsert, batch upsert, search, delete, get, list, stats, index creation, optimization, export | | **Quantum Operations** | 6 | Compress, decompress, hybrid similarity, space optimization, feature mapping, kernel similarity | | **Table Operations** | 8 | Full NoSQL database with MongoDB-style queries | | **File Operations** | 7 | Upload, download, list, delete, metadata, presigned URLs | | **Project Operations** | 7 | Create, manage, and monitor ZeroDB projects | | **Event Operations** | 5 | Event streaming and pub/sub | | **RLHF Operations** | 10 | Reinforcement learning from human feedback tracking | | **Admin Operations** | 5 | System stats, health checks, optimization | #### Technical Highlights | Feature | Details | |---------|---------| | **Async Support** | Full async/await support using httpx ⚑ | | **Retry Logic** | Exponential backoff retry logic πŸ”„ | | **Rate Limiting** | Intelligent rate limiting 🎯 | | **Error Handling** | Comprehensive with custom exception classes βœ… | | **Test Coverage** | 90%+ test coverage πŸ§ͺ | | **Python Version** | Python 3.9+ support 🐍 | > **Coming Soon:** Agent API support in Python SDK (currently available in TypeScript and Go) --- ### 🎯 **TypeScript SDK v2.3.0** - Full AINative SDK **Installation:** `npm install @ainative/zerodb-mcp-client` The TypeScript SDK delivers the complete AINative platform with **60+ ZeroDB operations** AND **26 Agent API operations** with full type safety and seamless integration with Node.js, React, Next.js, and other frameworks. #### πŸ†• Agent API Support (26 Operations) The SDK now includes comprehensive Agent API support across 5 modules: **Agent Orchestration (5 operations):** ```typescript import { ZeroDBClient } from '@ainative/zerodb-mcp-client'; const client = new ZeroDBClient({ apiKey: 'your_api_key', baseURL: 'https://api.ainative.studio' }); // Create agent tasks const task = await client.agentOrchestration.createTask({ agent_id: 'backend-agent', task_type: 'code_review', description: 'Review pull request #42', priority: 'high' }); // List tasks const tasks = await client.agentOrchestration.listTasks({ agent_id: 'backend-agent', status: 'completed' }); // Get task status const status = await client.agentOrchestration.getTaskStatus('task_123'); // Execute tasks const result = await client.agentOrchestration.executeTask({ task_id: 'task_123', params: { verbose: true } }); // Create task sequences const sequence = await client.agentOrchestration.createTaskSequence({ name: 'Build Pipeline', tasks: ['task_1', 'task_2', 'task_3'] }); ``` **Agent Coordination (3 operations):** ```typescript // Send messages between agents await client.agentCoordination.sendMessage({ from_agent: 'coordinator', to_agent: 'backend', message: 'Start code review', message_type: 'task_assignment' }); // Distribute tasks across agents const distribution = await client.agentCoordination.distributeTasks({ tasks: ['task_1', 'task_2', 'task_3'], agents: ['backend', 'frontend', 'database'], strategy: 'load_balanced' }); // Get agent workload statistics const stats = await client.agentCoordination.getWorkloadStats({ agent_id: 'backend' }); ``` **Agent State Management (4 operations):** ```typescript // Get current agent state const state = await client.agentState.getState({ agent_id: 'backend' }); // Create state checkpoints const checkpoint = await client.agentState.createCheckpoint({ agent_id: 'backend', name: 'pre_deployment', data: { config: { debug: true } } }); // Restore from checkpoint await client.agentState.restoreCheckpoint({ checkpoint_id: 'cp_123', agent_id: 'backend' }); // List available checkpoints const checkpoints = await client.agentState.listCheckpoints({ agent_id: 'backend' }); ``` **Agent Swarm Management (11 operations):** ```typescript // Create agent swarms const swarm = await client.agentSwarm.createSwarm({ name: 'Development Team', project_id: 'proj_123', config: { max_agents: 10, timeout: 300 } }); // Scale swarms await client.agentSwarm.scaleSwarm({ swarm_id: 'swarm_123', count: 5 }); // Get swarm analytics const analytics = await client.agentSwarm.getSwarmAnalytics({ swarm_id: 'swarm_123', period: '7d' }); // Start/stop swarms await client.agentSwarm.startSwarm({ swarm_id: 'swarm_123' }); await client.agentSwarm.stopSwarm({ swarm_id: 'swarm_123' }); // Plus: listSwarms, deleteSwarm, getAgentTypes, getStatus, updateSwarmConfig ``` **Agent Learning (3 operations):** ```typescript // Submit feedback await client.agentLearning.submitFeedback({ agent_id: 'backend', interaction_id: 'int_123', rating: 4.5, comments: 'Great code review' }); // Get performance metrics const metrics = await client.agentLearning.getPerformanceMetrics({ agent_id: 'backend', period: '7d' }); // Compare agent performance const comparison = await client.agentLearning.compareAgents({ agents: ['backend', 'frontend', 'database'], metric: 'success_rate' }); ``` #### ZeroDB Operations All 60+ ZeroDB operations are included with full type safety: ```typescript // Vector search const results = await client.vectors.search({ project_id: 'your-project-uuid', query_vector: embedding, limit: 10, threshold: 0.7 }); // Quantum operations const compressed = await client.quantum.compress({ project_id: 'project-uuid', embedding: largeVector, compression_ratio: 0.6 }); // NoSQL tables await client.tables.insertRows({ project_id: 'project-uuid', table_name: 'users', rows: [{ name: 'John', email: 'john@example.com' }] }); ``` #### Enterprise Features | Feature | Description | |---------|-------------| | **Complete Type Safety** | Full TypeScript definitions for all 86+ operations βœ… | | **Error Handling** | Custom error classes (AuthenticationError, RateLimitError, ValidationError) 🚨 | | **Retry Logic** | Built-in exponential backoff for transient failures πŸ”„ | | **Authentication** | Support for both API keys and JWT tokens πŸ” | | **Framework Support** | Works seamlessly with React, Next.js, Vue, Angular βš›οΈ | | **Validation** | Input validation for vectors, UUIDs, and data types βœ”οΈ | --- ### ⚑ **Go SDK** - High-Performance with Full Agent Support **Installation:** `go get github.com/ainative/go-sdk` Built for performance and enterprise reliability, the Go SDK delivers **50%+ faster** execution compared to Python with **full Agent API support** and advanced concurrency features. #### Agent API Support (26 Operations) ```go import "github.com/ainative/go-sdk/ainative" // Initialize client client := ainative.NewClient(ainative.Config{ APIKey: "your_api_key", MaxConnections: 100, Timeout: 30 * time.Second, }) // Agent Orchestration task, err := client.AgentOrchestration.CreateTask(ctx, ainative.CreateTaskRequest{ AgentID: "backend-agent", TaskType: "code_review", Description: "Review pull request #42", Priority: "high", }) // Agent Coordination message, err := client.AgentCoordination.SendMessage(ctx, ainative.SendMessageRequest{ FromAgent: "coordinator", ToAgent: "backend", Message: "Start code review", MessageType: "task_assignment", }) // Agent State Management state, err := client.AgentState.GetState(ctx, ainative.GetStateRequest{ AgentID: "backend", }) // Agent Swarm swarm, err := client.AgentSwarm.CreateSwarm(ctx, ainative.CreateSwarmRequest{ Name: "Development Team", ProjectID: "proj_123", Config: map[string]interface{}{"max_agents": 10}, }) // Agent Learning feedback, err := client.AgentLearning.SubmitFeedback(ctx, ainative.SubmitFeedbackRequest{ AgentID: "backend", InteractionID: "int_123", Rating: 4.5, Comments: "Great performance", }) ``` #### ZeroDB Operations ```go // High-performance batch vector operations vectors := []ainative.Vector{ {Embedding: embedding1, Document: "Document 1"}, {Embedding: embedding2, Document: "Document 2"}, } result, err := client.Vectors.BatchUpsert(ctx, vectors) if err != nil { log.Fatal(err) } ``` #### Enterprise-Grade Features | Feature | Benefit | |---------|---------| | **High-Performance HTTP Client** | Advanced connection pooling for concurrent operations πŸš€ | | **Circuit Breaker** | Prevent cascade failures with intelligent circuit breaking ⚑ | | **Exponential Backoff** | Retry logic with jitter for optimal reliability πŸ”„ | | **OpenTelemetry Integration** | Built-in metrics and distributed tracing πŸ“Š | | **Context Propagation** | Full context support throughout the SDK πŸ”— | | **Type Safety** | Comprehensive struct definitions for all operations βœ… | | **Security Patching** | Active CVE monitoring and rapid security updates πŸ”’ | #### Performance Benchmarks | Metric | Value | Impact | |--------|-------|--------| | **Batch Operations** | 50%+ faster than Python | **High Performance** ⚑ | | **Concurrent Connections** | 100+ with efficient pooling | **Scalable** πŸ“ˆ | | **Operation Overhead** | Sub-millisecond | **Low Latency** 🎯 | | **Reliability** | Enterprise-grade retry/timeout | **Production Ready** βœ… | --- ## Common Use Cases ### 1. **Multi-Agent Systems** Build coordinated agent teams that work together: **TypeScript:** ```typescript // Create a development swarm const swarm = await client.agentSwarm.createSwarm({ name: 'Dev Team', project_id: 'proj_123', config: { max_agents: 5 } }); // Distribute tasks across the swarm await client.agentCoordination.distributeTasks({ tasks: ['review_pr', 'write_tests', 'deploy'], agents: ['backend', 'frontend', 'devops'], strategy: 'load_balanced' }); // Monitor agent performance const metrics = await client.agentLearning.getPerformanceMetrics({ agent_id: 'backend', period: '7d' }); ``` **Go:** ```go swarm, err := client.AgentSwarm.CreateSwarm(ctx, ainative.CreateSwarmRequest{ Name: "Dev Team", ProjectID: "proj_123", Config: map[string]interface{}{"max_agents": 5}, }) distribution, err := client.AgentCoordination.DistributeTasks(ctx, ainative.DistributeTasksRequest{ Tasks: []string{"review_pr", "write_tests", "deploy"}, Agents: []string{"backend", "frontend", "devops"}, Strategy: "load_balanced", }) ``` ### 2. **Semantic Search with Agent Memory** Combine ZeroDB and Agent APIs for intelligent systems: **TypeScript:** ```typescript // Store agent interaction in ZeroDB const result = await client.embeddings.embed_and_store({ project_id: 'agent-memory', text: agent_response, metadata: { agent_id: 'customer-support', interaction_id: 'int_123' } }); // Submit agent feedback await client.agentLearning.submitFeedback({ agent_id: 'customer-support', interaction_id: 'int_123', rating: 4.5 }); // Search agent memory const memory = await client.embeddings.semantic_search({ project_id: 'agent-memory', query: 'How did we handle payment issues?', limit: 5 }); ``` ### 3. **Agent State Checkpointing** Save and restore agent states for reliability: **TypeScript:** ```typescript // Create checkpoint before risky operation const checkpoint = await client.agentState.createCheckpoint({ agent_id: 'production-bot', name: 'pre_deployment', data: currentState }); // If deployment fails, restore if (deploymentFailed) { await client.agentState.restoreCheckpoint({ checkpoint_id: checkpoint.id, agent_id: 'production-bot' }); } ``` --- ## MCP Server Integration **Separate from SDKs:** The **ainative-zerodb-mcp-server** is a Model Context Protocol server that allows AI assistants like Claude to interact with ZeroDB: ```bash # Install MCP server (separate package) npm install -g ainative-zerodb-mcp-server ``` ```json { "mcpServers": { "zerodb": { "command": "npx", "args": ["-y", "ainative-zerodb-mcp-server"], "env": { "ZERODB_PROJECT_ID": "your-project-id", "ZERODB_USERNAME": "your-username", "ZERODB_PASSWORD": "your-password" } } } } ``` This allows Claude Code, Claude Desktop, and other MCP clients to: - Store and retrieve agent memory via ZeroDB - Perform semantic searches - Manage vector databases - Track RLHF feedback **Note:** The MCP server focuses on ZeroDB operations. For programmatic Agent API access, use the SDKs directly. --- ## Getting Started ### Python Developers (ZeroDB Operations) ```bash pip install zerodb-mcp ``` πŸ“š [Python SDK Documentation](https://docs.ainative.studio/sdk/python) πŸ“¦ [PyPI Package](https://pypi.org/project/zerodb-mcp/) ### TypeScript/JavaScript Developers (ZeroDB + Agent APIs) ```bash npm install @ainative/zerodb-mcp-client ``` πŸ“š [TypeScript SDK Documentation](https://docs.ainative.studio/sdk/typescript) πŸ“¦ [NPM Package](https://www.npmjs.com/package/@ainative/zerodb-mcp-client) ### Go Developers (ZeroDB + Agent APIs) ```bash go get github.com/ainative/go-sdk ``` πŸ“š [Go SDK Documentation](https://docs.ainative.studio/sdk/go) πŸ’» [GitHub Repository](https://github.com/AINative-Studio/zerodb-go-sdk) --- ## Migration Guide for Existing Users ### Python SDK v1.0.x β†’ v1.1.0 The embeddings module is now exposed on the client: ```python # Before (manual embedding required) import openai embedding = openai.Embedding.create(input=text) await client.vectors.upsert(embedding=embedding['data'][0]['embedding']) # After (FREE embeddings included!) result = await client.embeddings.embed_and_store( project_id=project_id, text=text ) ``` ### TypeScript SDK v2.2.x β†’ v2.3.0 **Major Update:** 26 Agent API operations added! ```typescript // NEW: Agent operations now available const task = await client.agentOrchestration.createTask({ agent_id: 'backend', task_type: 'code_review', description: 'Review PR #42' }); // All existing ZeroDB operations remain unchanged const results = await client.vectors.search({ ... }); ``` ### Go SDK First production release with full Agent API support - no migration needed for new users. --- ## SDK Comparison | Feature | Python SDK | TypeScript SDK | Go SDK | |---------|:----------:|:--------------:|:------:| | **ZeroDB Operations** | βœ… 60+ | βœ… 60+ | βœ… 60+ | | **Agent APIs** | πŸ”œ Coming Soon | βœ… 26 ops | βœ… 26 ops | | **FREE Embeddings** | βœ… | ❌ | ❌ | | **Type Safety** | βœ… Pydantic | βœ… TypeScript | βœ… Go types | | **Async Support** | βœ… async/await | βœ… Promises | βœ… Goroutines | | **Performance** | Good | Good | **50%+ faster** πŸš€ | | **Frameworks** | Python 3.9+ | Node, React, Next | Any Go app | | **Test Coverage** | 90%+ | 85%+ | 85%+ | --- ## What's Next? We're committed to continuous improvement across all SDKs: **Q1 2026 Roadmap:** - βœ… Agent API support in Python SDK - WebSocket support for real-time events (Python & TypeScript) - Streaming response support for large results - Enhanced caching mechanisms - CLI tools for common operations - Additional quantum algorithms - Multi-region support --- ## Community & Support Join our growing community of AI developers: - πŸ“– **Documentation**: [docs.ainative.studio](https://docs.ainative.studio) - πŸ’¬ **Discord**: [AINative Studio Community](https://discord.gg/ainative) - πŸ› **GitHub Issues**: [Python](https://github.com/ainative/zerodb-mcp-python/issues) | [TypeScript](https://github.com/ainative-studio/zerodb-mcp-client/issues) | [Go](https://github.com/AINative-Studio/zerodb-go-sdk/issues) - πŸ“§ **Email**: support@ainative.studio - 🐦 **Twitter/X**: [@ainativestudio](https://x.com/ainativestudio) --- ## Try It Today All SDKs are **production-ready** and available now: ```bash # Python (ZeroDB + FREE embeddings) pip install zerodb-mcp # TypeScript (ZeroDB + Agent APIs) npm install @ainative/zerodb-mcp-client # Go (ZeroDB + Agent APIs) go get github.com/ainative/go-sdk ``` Get your **FREE API key** at [ainative.studio/signup](https://www.ainative.studio/signup) and start building AI-native applications today. **Free Tier Includes:** - 1 million vector operations/month - 1 million agent API calls/month - 10 GB vector storage - Unlimited embedding generation (Python only) - Full access to all 86+ operations - Community support Ready to scale? Check out our [pricing plans](https://www.ainative.studio/pricing) for production workloads. --- *Built with ❀️ by the AINative Studio team. Questions? Reach out at support@ainative.studio*
AI DevelopmentSDKsPythonTypeScriptGoModel Context ProtocolMCP ServersZeroDB

Check your site's AX Score

Free scan, 6 categories, under 60 seconds. See how your site ranks on the agentic web.

Run a free audit β†’