langchain-ai/deepagents · error · ValueError

response_schema exceeds {_SCHEMA_MAX_BYTES} byte limit ({len

Error message

response_schema exceeds {_SCHEMA_MAX_BYTES} byte limit ({len(serialized)} bytes)

What it means

`_validate_response_schema` serializes the caller's JSON Schema with `json.dumps` and enforces a maximum size of `_SCHEMA_MAX_BYTES`. Oversized schemas are rejected with `ValueError` before being handed to the subagent runtime, keeping prompts and provider payloads bounded.

Source

Thrown at libs/partners/quickjs/langchain_quickjs/_subagent.py:283

        "phase": "complete",
        "id": subagent_id,
        "duration_ms": int((time.monotonic() - started_at) * 1000),
    }
    if eval_id is not None:
        complete_event["eval_id"] = eval_id
    _emit_subagent_event(stream_writer, complete_event)
    return output


def _validate_response_schema(schema: dict[str, Any]) -> None:
    """Reject schemas that exceed size, depth, or property-count limits."""
    serialized = json.dumps(schema)
    if len(serialized) > _SCHEMA_MAX_BYTES:
        msg = (
            f"response_schema exceeds {_SCHEMA_MAX_BYTES}"
            f" byte limit ({len(serialized)} bytes)"
        )
        raise ValueError(msg)

    def _check(node: dict[str, Any], depth: int, prop_count: list[int]) -> None:
        if depth > _SCHEMA_MAX_DEPTH:
            msg = (
                f"response_schema exceeds maximum nesting depth of {_SCHEMA_MAX_DEPTH}"
            )
            raise ValueError(msg)
        props = node.get("properties")
        if isinstance(props, dict):
            prop_count[0] += len(props)
            if prop_count[0] > _SCHEMA_MAX_PROPERTIES:
                msg = (
                    "response_schema exceeds maximum of"
                    f" {_SCHEMA_MAX_PROPERTIES} properties"
                )
                raise ValueError(msg)
            for value in props.values():
                if isinstance(value, dict):

View on GitHub (pinned to a1af029e6e)

Solutions

1. Shrink the schema to the minimal structure you need — describe only the fields you will consume.
2. Move large shared definitions out (or reference them by name) and deduplicate repeated subtrees.
3. Split the extraction into multiple `task()` calls with smaller schemas.
4. If the limit is genuinely too small, check `_SCHEMA_MAX_BYTES` for the current bound and raise a targeted issue rather than bypassing validation.

Example fix

// before
schema = load_entire_openapi_schema()  # ~200 KB
task(prompt="extract user", response_schema=schema)

// after
schema = {"type": "object", "properties": {"name": {"type": "string"}, "email": {"type": "string"}}}
task(prompt="extract user", response_schema=schema)

import json
serialized = json.dumps(schema)
if len(serialized.encode()) > _SCHEMA_MAX_BYTES:
    raise ValueError("response_schema too large; trim to required fields")

try:
    task(prompt=prompt, response_schema=schema)
except ValueError as e:
    if "byte limit" in str(e):
        task(prompt=prompt, response_schema=minify_schema(schema))

When it happens

Trigger: Passing a `response_schema` larger than `_SCHEMA_MAX_BYTES` to `task()` (via `call_subagent_task_tool`), typically an inlined mega-schema, a schema with embedded data/examples, or a duplicated deeply-nested schema object.

Common situations: Generating schemas programmatically that balloon with `$defs` repetition; pasting a full OpenAPI-derived schema where a small response shape was intended; embedding sample payloads inside the schema.

Understand the failure class

Background: Schema validation failed / invalid input schema: payload rejected because its shape doesn't match the expected schema — this error's family across 28 libraries.

Related errors

• response_schema exceeds maximum nesting depth of {_SCHEMA_MA
• response_schema exceeds maximum of {_SCHEMA_MAX_PROPERTIES}
• Could not parse embedded resource block. Block expected eith
• {question_type} question {question_text!r} must not define '
• Unknown external event kind: {self.kind!r}

AI-assisted analysis of langchain-ai/deepagents@a1af029e6e (2026-08-29). Data as JSON: /api/errors/b81c8dd067748c0b. Report an issue: GitHub.