Core Concepts

Six concepts cover everything CubePi does. Read this page once, then the rest of the docs become a lookup table.

Agent

Agent is the stateful façade: you construct it with a provider, a model, optional tools, and optional middleware/checkpointer. You drive it through three methods:

• await agent.prompt(message) — start a new turn from a user message.
• await agent.resume() — continue from the last persisted message (used with a checkpointer).
• agent.steer(message) / agent.follow_up(message) — queue a message mid-flight or after the current run.

The agent owns an AgentState (system prompt, tools, model, message history, pending tool calls, streaming flag) and a list of subscribers:

unsubscribe = agent.subscribe(my_listener)
# ...
unsubscribe()

Subscribers receive every AgentEvent the loop emits. They can be sync or async.

Tool

An AgentTool is a name + description + Pydantic parameter model + async execute function:

from pydantic import BaseModel
from cubepi import AgentTool, AgentToolResult, TextContent

class SearchParams(BaseModel):
    query: str
    limit: int = 10

async def execute(tool_call_id, params: SearchParams, *, signal=None, on_update=None):
    # do work; respect `signal` if you can be aborted
    return AgentToolResult(content=[TextContent(text=f"…")])

search = AgentTool(
    name="search",
    description="Search the corpus",
    parameters=SearchParams,
    execute=execute,
)

The Pydantic schema is auto-converted to JSON Schema and passed to the model. Arg parsing, error wrapping, and parallel execution are handled by the framework. See Tool Use for execution_mode, on_update (incremental progress), and terminate (end the turn from a tool).

Provider

A Provider is anything matching this Protocol:

class Provider(Protocol):
    async def stream(
        self,
        model: Model,
        messages: list[Message],
        *,
        system_prompt: str = "",
        tools: list[ToolDefinition] | None = None,
        options: StreamOptions | None = None,
    ) -> MessageStream: ...

It returns a MessageStream — a single async iterator that yields StreamEvents and exposes the final AssistantMessage via await stream.result(). Built-in providers:

• AnthropicProvider — Claude (Messages API, with thinking, caching, tool use).
• OpenAIProvider — GPT family (Chat Completions API).
• OpenAIResponsesProvider — GPT family (Responses API, server-side state).
• FauxProvider — deterministic test double (no network).

Write your own by implementing one method. See Providers / Custom.

Stream and events

Streams and events are two layers:

• Provider streams — MessageStream yields provider events: start, text_start, text_delta, text_end, thinking_*, toolcall_*, done, error. This is the raw token stream.
• Agent events — what agent.subscribe(...) receives. Eleven types covering the entire loop: agent_start, agent_end, turn_start, turn_end, message_start, message_update, message_end, tool_execution_start, tool_execution_update, tool_execution_end. message_update wraps the provider event inside event.stream_event.

Subscribe to agent events for UI; for low-level token routing dig into event.stream_event. See Streaming Events.

Middleware

Middleware is a class with up to seven typed hooks:

Hook	When it runs	Composition rule
transform_context	Before each model call, on the message list	Chained — each receives previous result
convert_to_llm	Right before serialisation to the provider	Last implementation wins
transform_system_prompt	Before each model call, on the system prompt	Chained
before_tool_call	Per tool call, after arg validation	First block=True short-circuits
after_tool_call	Per tool call, after execute	Later override earlier
after_model_response	After the assistant message lands	Returns a TurnAction controlling flow
should_stop_after_turn	At each turn boundary	Any True stops

Pass middleware as a list to Agent(middleware=[...]). See Middleware → Composition.

Checkpointer

A Checkpointer is anything matching:

class Checkpointer(Protocol):
    async def load(self, thread_id: str) -> CheckpointData | None: ...
    async def append(self, thread_id: str, messages: list[Message]) -> None: ...
    async def save_extra(self, thread_id: str, extra: dict) -> None: ...

Bind one to an agent with Agent(checkpointer=cp, thread_id="…") and the loop will append each new message as it lands, restoring history on the first prompt(). Built-in backends: MemoryCheckpointer, SQLiteCheckpointer, PostgresCheckpointer. See Checkpointing → SQLite.

Tracer (optional)

Tracer produces OpenTelemetry spans aligned with the GenAI Semantic Conventions, so any OTLP backend (Jaeger, Tempo, Honeycomb, Datadog, AWS X-Ray, …) can ingest agent runs without custom instrumentation. Install the extra:

pip install "cubepi[tracing]"           # OTel SDK
pip install "cubepi[tracing-otlp]"      # + OTLP/HTTP exporter

then wrap your agent in an async with:

from cubepi.tracing import Tracer
from cubepi.tracing.exporters import JsonlSpanExporter

async with (
    Tracer(
        service_name="my-bot",
        agent_name="assistant",
        exporters=[JsonlSpanExporter(directory="./cubepi-traces")],
    ) as tracer,
    tracer.attached(agent),
):
    await agent.prompt("…")

Each run emits an invoke_agent root span containing one cubepi.turn per LLM round-trip, plus chat (CLIENT) and execute_tool children. By default no prompt content or model output is recorded — opt in with Tracer(record_content=True) and a redact callback for PII. Pair with Meter(...) for token / duration / TTFC histograms. Full guide: Tracing → Overview.

Putting it together

User code
   │
   ▼
┌──────────────────────────────────────────┐
│ Agent                                     │
│  ├─ AgentState (messages, tools, …)       │
│  ├─ Middleware ── compose_middleware()    │
│  ├─ Checkpointer ── append on message_end │
│  └─ run_agent_loop  ◀──── the actual loop │
│       │                                   │
│       ▼                                   │
│  Provider.stream() → MessageStream        │
│       │                                   │
│       └─ events → emit → subscribers      │
└──────────────────────────────────────────┘

That diagram is the whole framework. The rest of this site is just the details.