langchain-ai/langchain · error · ValueError

Trying to deserialize something that cannot be deserialized

Error message

Trying to deserialize something that cannot be deserialized in current version of langchain-core: {mapping_key}.

What it means

Raised when a serialized object's namespace root appears in DISALLOW_LOAD_FROM_PATH: namespaces (like the legacy 'langchain' root) whose modules cannot be imported by path in current langchain-core. Even if the namespace passes validation, these legacy roots deliberately refuse path-based imports, so deserialization fails.

Source

Thrown at libs/core/langchain_core/load/load.py:538

            if (
                namespace[0] not in self.valid_namespaces
                # The root namespace ["langchain"] is not a valid identifier.
                or namespace == ["langchain"]
            ):
                msg = f"Invalid namespace: {value}"
                raise ValueError(msg)
            # Determine explicit import path
            if mapping_key in self.import_mappings:
                import_path = self.import_mappings[mapping_key]
                # Split into module and name
                import_dir, name = import_path[:-1], import_path[-1]
            elif namespace[0] in DISALLOW_LOAD_FROM_PATH:
                msg = (
                    "Trying to deserialize something that cannot "
                    "be deserialized in current version of langchain-core: "
                    f"{mapping_key}."
                )
                raise ValueError(msg)
            else:
                # Otherwise, treat namespace as path.
                import_dir = namespace

            # Validate import path is in trusted namespaces before importing
            if import_dir[0] not in self.valid_namespaces:
                msg = f"Invalid namespace: {value}"
                raise ValueError(msg)

            # We don't need to recurse on kwargs
            # as json.loads will do that for us.
            kwargs = value.get("kwargs", {})

            # Run the init_validator (e.g., jinja2 blocking) before importing
            # to fail fast on security violations.
            if self.init_validator is not None:
                self.init_validator(mapping_key, kwargs)

View on GitHub (pinned to e32fa9a52e)

Solutions

1. Rebuild the object in the current version instead of loading the legacy payload
2. Provide an import mapping redirect: loads(text, import_mappings={('langchain','chains','LLMChain'): ('langchain_classic','chains','llm','LLMChain')})
3. Check langchain_core.load.mapping for existing compat mappings before adding your own

Example fix

# before
obj = loads(legacy_payload)
# after
obj = loads(legacy_payload, import_mappings={('langchain','chains','LLMChain'): ('langchain_classic','chains','llm','LLMChain')})

from langchain_core.load.load import DISALLOW_LOAD_FROM_PATH
import json
payload = json.loads(text)
banned = {n['id'][0] for n in iter_constructor_nodes(payload)} & set(DISALLOW_LOAD_FROM_PATH)
if banned:
    raise ValueError(f'legacy roots {banned} need import_mappings redirects')

try:
    obj = loads(text)
except ValueError as e:
    if 'cannot be deserialized' in str(e):
        obj = loads(text, import_mappings=LEGACY_REDIRECTS)
    else:
        raise

When it happens

Trigger: Loading old payloads with ids like ['langchain', 'chains', 'LLMChain'] where no import_mapping entry covers the class — the fallback import_dir = namespace path is disallowed. Typically seen with pre-0.1 serialized chains not present in the compat import_mappings.

Common situations: Restoring chains serialized under langchain <=0.0.x; loading exports from old LangSmith runs; class was removed or renamed so no mapping redirects it.

Related errors

• Invalid namespace: {value}
• Tracing using LangChainTracerV1 is no longer supported. Plea
• Jinja2 templates are not allowed during deserialization for
• allowed_objects must contain Serializable subclasses.
• Trying to load an object that doesn't implement serializatio

AI-assisted analysis of langchain-ai/langchain@e32fa9a52e (2026-08-14). Data as JSON: /api/errors/eae6a72cfbdf70b9. Report an issue: GitHub.