langflow-ai/langflow · error · HTTPException

str(exc)

Error message

str(exc)

What it means

A 400 whose detail is the message of a ValueError raised by FolderSource.validate_config() on the folder-ingest endpoint. Validation happens up-front so invalid folder configurations fail before any background job is spawned. The detail string is produced by the FolderSource implementation (path allow-list, existence, recursion settings, security settings applied via _apply_folder_source_security_settings).

Source

Thrown at src/backend/base/langflow/api/v1/knowledge_bases.py:1273

        # Build + validate the folder source up-front so invalid
        # configurations surface as a 4xx response before a background
        # job is spawned.
        source_config: dict[str, Any] = {
            "path": payload.path,
            "recursive": payload.recursive,
        }
        if payload.extensions is not None:
            source_config["extensions"] = payload.extensions
        if per_file_user_metadata:
            source_config["per_file_metadata"] = per_file_user_metadata
        source_config = _apply_folder_source_security_settings(source_config)

        folder_source = FolderSource(user_id=current_user.id, source_config=source_config)
        try:
            await folder_source.validate_config()
        except ValueError as exc:
            raise HTTPException(status_code=400, detail=str(exc)) from exc

        job_service = get_job_service()
        job_id = uuid.uuid4()

        await job_service.create_job(
            job_id=job_id,
            flow_id=job_id,
            job_type=JobType.INGESTION,
            asset_id=asset_id,
            asset_type="knowledge_base",
            user_id=current_user.id,
        )

        task_service = get_task_service()
        await task_service.fire_and_forget_task(
            job_service.execute_with_status,
            job_id=job_id,
            run_coro_func=KBIngestionHelper.perform_ingestion,

View on GitHub (pinned to 976ec789d2)

Solutions

1. Read the returned detail message — it names the exact validation failure (missing path, not a directory, disallowed root).
2. Ensure payload.path exists on the server and is readable by the backend process.
3. If the operator allow-lists ingest roots, move data under an allowed root or ask the operator to extend the allow-list (then restart).
4. For containerized deployments, mount the data directory into the container and use the in-container path.

Example fix

# before
payload = {"path": "/home/me/data", "recursive": True}

# after (docker: data mounted at /data)
payload = {"path": "/data", "recursive": True}

import os

def folder_payload_valid(path: str, allowed_roots: list[str]) -> bool:
    expanded = os.path.abspath(os.path.expanduser(path))
    return os.path.isdir(expanded) and any(expanded.startswith(r) for r in allowed_roots)

try:
    resp = await client.post(folder_url, json=payload)
except HTTPStatusError as e:
    if e.response.status_code == 400:
        surface_error(e.response.json()["detail"])  # FolderSource message names the field

When it happens

Trigger: POST /api/v1/knowledge_bases/{kb_name}/ingest/folder with a payload.path that does not exist, is not a directory, is outside the operator-configured allow-list of ingestible roots, or otherwise fails FolderSource validation (e.g. path denied by server-side security settings).

Common situations: Running the backend in a container so paths like /home/user/data on the host are absent; forgetting that ~ expands to the server's home, not the client's; the operator restricting folder ingestion to specific roots via security settings.

Related errors

• Failed to install MCP
• Invalid path
• Provider '{provider}' is not configured. Available providers
• Could not extract class name from code
• Inputs and outputs have overlapping names: {overlap}

AI-assisted analysis of langflow-ai/langflow@976ec789d2 (2026-08-14). Data as JSON: /api/errors/71f36d0441fc6164. Report an issue: GitHub.