API Parity

Complete mapping between TypeScript and Python APIs.

Naming Convention

All TypeScript camelCase names map to Python snake_case:

TypeScript	Python
`defineAgent()`	`define_agent()`
`defineTool()`	`define_tool()`
`createLLM()`	`create_llm()`
`createMockLLM()`	`create_mock_llm()`
`createTestHarness()`	`create_test_harness()`
`systemPrompt`	`system_prompt`
`maxSteps`	`max_steps`
`maxTokens`	`max_tokens`
`toolCalls`	`tool_calls`
`runId`	`run_id`

Schema Validation

TypeScript	Python
Zod (`z.object({...})`)	Pydantic (`BaseModel`)
`z.string()`	`str`
`z.number()`	`float` / `int`
`z.boolean()`	`bool`
`z.array(z.string())`	`list[str]`
`z.optional()`	`Optional[T]`
`.describe('...')`	`Field(description='...')`

Complete API Map

Core

TypeScript	Python
`Agent`	`Agent`
`defineAgent()`	`define_agent()`
`defineTool()`	`define_tool()`
`createLLM()`	`create_llm()`

State

TypeScript	Python
`InMemoryStateStore`	`InMemoryStateStore`
`RedisStateStore`	`RedisStateStore`

Memory

TypeScript	Python
`InMemoryMemoryStore`	`InMemoryMemoryStore`
`PgVectorMemoryStore`	`PgVectorMemoryStore`
`retrieveMemories()`	`retrieve_memories()`
`consolidateMemories()`	`consolidate_memories()`

Governor

TypeScript	Python
`BudgetGovernor`	`BudgetGovernor`
`COST_MODELS`	`COST_MODELS`
`AutonomyPolicy`	`AutonomyPolicy`

Multi-Agent

TypeScript	Python
`Coordinator`	`Coordinator`
`pipeline()`	`pipeline()`
`parallel()`	`parallel()`
`supervisor()`	`supervisor()`
`debate()`	`debate()`

Testing

TypeScript	Python
`createMockLLM()`	`create_mock_llm()`
`createTestHarness()`	`create_test_harness()`

Observability

TypeScript	Python
`getInstrumentation()`	`get_instrumentation()`
`resetInstrumentation()`	`reset_instrumentation()`

Next Steps

Differences — Python-specific behavior

Python Quickstart

Get started with AgentForge in Python — install, configure, and run your first agent.

Python Differences

Key differences between the TypeScript and Python implementations.

On this page

Naming Convention Schema Validation Complete API Map Core State Memory Governor Multi-Agent Testing Observability Next Steps