Documentation Index
Fetch the complete documentation index at: https://docs.vectorshift.ai/llms.txt
Use this file to discover all available pages before exploring further.
Lifecycle
new
Pipeline.new(name: str, nodes: Optional[list[Node]] = None) -> Pipeline
The name of the pipeline.
nodes
Optional[list[Node]]
default:"None"
List of nodes to include in the pipeline.
save
Pipeline.save(deploy: bool = True, bump: Optional[BumpLevel] = None, description: Optional[str] = None)
When True, triggers deployment semantics (versioning/merge).
bump
Optional[BumpLevel]
default:"None"
Version bump level: “major”, “minor”, or “patch”. See BumpLevel. description
Optional[str]
default:"None"
Updates the pipeline description and is used as the version changelog.
dict: A dictionary containing the status of the save operation.
If the pipeline update fails.
fetch
Pipeline.fetch(id: Optional[str] = None, name: Optional[str] = None, branch_id: Optional[str] = None, version: Optional[str] = None, username: Optional[str] = None, org_name: Optional[str] = None) -> Pipeline
branch_id and version only apply when fetching by id.
Parameters
id
Optional[str]
default:"None"
The unique identifier of the pipeline to fetch.
name
Optional[str]
default:"None"
The name of the pipeline to fetch. When multiple pipelines share the same name, the most recently modified one is returned.
branch_id
Optional[str]
default:"None"
Fetch the pipeline for a specific branch (id-fetch only).
version
Optional[str]
default:"None"
Semantic version to fetch (e.g. "1.2.0"), or "latest" (id-fetch only).
username
Optional[str]
default:"None"
The username of the pipeline owner.
org_name
Optional[str]
default:"None"
The organization name of the pipeline owner.
Pipeline: The fetched Pipeline instance.
If neither id nor name is provided.
list
Pipeline.list(folder_id: Optional[str] = None, include_shared: Optional[bool] = None, verbose: Optional[bool] = None, offset: Optional[int] = None, limit: Optional[int] = None) -> "list[Pipeline]"
folder_id
Optional[str]
default:"None"
Filter pipelines by folder.
include_shared
Optional[bool]
default:"None"
Include shared pipelines.
verbose
Optional[bool]
default:"None"
Include full pipeline details.
offset
Optional[int]
default:"None"
Pagination offset.
limit
Optional[int]
default:"None"
Maximum number of pipelines to return.
A list of Pipeline instances.
delete
Deletes an existing pipeline.
Returns
dict: A dictionary containing the status of the deletion operation.
If the pipeline couldn’t be deleted.
duplicate
Pipeline.duplicate(name: Optional[str] = None, description: Optional[str] = None, folder_id: Optional[str] = None) -> Pipeline
name
Optional[str]
default:"None"
Name for the duplicate. Auto-generated if omitted.
description
Optional[str]
default:"None"
Description for the duplicate.
folder_id
Optional[str]
default:"None"
Folder to place the duplicate in.
A new Pipeline instance for the duplicated pipeline.
Building
add
Pipeline.add -> NodeAdder # property
inp = pipeline.add.input(name="query", input_type="string")
llm = pipeline.add.llm(provider="openai", model="gpt-4o", prompt=inp.text)
add_batch
Pipeline.add_batch -> BatchNodeAdder # property
execution_mode='batch'.
Usage
llm = pipeline.add_batch.llm(provider="openai", model="gpt-4o", prompt=inp.text)
# equivalent to: pipeline.add.llm(..., execution_mode="batch")
add_node
Pipeline.add_node(node: Node) -> None
The node to add to the pipeline.
remove_node
Pipeline.remove_node(node_id: Optional[str] = None, node_name: Optional[str] = None) -> dict[str, Any]
node_id
Optional[str]
default:"None"
ID of the node to remove.
node_name
Optional[str]
default:"None"
Name of the node to remove.
\{"status": "success", "node_id": "\<resolved_node_id>"\}
Running
run
Pipeline.run(inputs: dict[str, Any], stream: bool = False, stream_all_outputs: bool = False, session_id: Optional[str] = None, node_input_overrides: Optional[dict[str, Any]] = None, send_intermediate_results: Optional[bool] = None) -> Union[dict[str, Any], Generator]
Dictionary of input nodes -> input values for the pipeline. eg: {“input_node”: “Hello, world!”}
Whether to stream the response. (Set true only when pipeline has an output node with a streaming llm input)
Whether to stream all outputs as they arrive.
session_id
Optional[str]
default:"None"
Optional session ID for run grouping and tracing. Groups multiple runs under the same session for analytics and observability. Note: this does not provide conversational memory to LLM nodes — each run’s LLM nodes only see the current inputs.
node_input_overrides
Optional[dict[str, Any]]
default:"None"
Optional per-node input overrides.
send_intermediate_results
Optional[bool]
default:"None"
Whether to send intermediate results (default: True on server).
returns
Union[dict[str, Any], Generator]
Union[dict[str, Any], Generator]: A dictionary containing pipeline outputs and run_id. If stream is True, returns a generator that yields response chunks.
If the pipeline execution fails.
bulk_run
Pipeline.bulk_run(inputs: list[dict[str, Any]]) -> list[dict[str, Any]]
inputs
list[dict[str, Any]]
required
List of dictionaries of input values for the pipeline.
A list of dictionaries containing the run_id and outputs for each set of inputs.
If the pipeline execution fails.
run_status
Pipeline.run_status(run_id: str) -> RunStatus
The run/task ID returned by :meth:start.
A :class:RunStatus dict with keys task_id, status, and
optionally error or result. See RunStatus. Background runs
start
Pipeline.start(inputs: dict[str, Any], session_id: Optional[str] = None, node_input_overrides: Optional[dict[str, Any]] = None, send_intermediate_results: Optional[bool] = None, webhook_url: Optional[str] = None, version: Optional[str] = None, trace_id: Optional[str] = None) -> RunHandler
Dictionary of input values for the pipeline.
session_id
Optional[str]
default:"None"
Optional session ID for run grouping and tracing. Groups multiple runs under the same session for analytics and observability. Note: this does not provide conversational memory to LLM nodes — each run’s LLM nodes only see the current inputs.
node_input_overrides
Optional[dict[str, Any]]
default:"None"
Optional per-node input overrides.
send_intermediate_results
Optional[bool]
default:"None"
Whether to send intermediate results.
webhook_url
Optional[str]
default:"None"
URL to call when the run completes.
version
Optional[str]
default:"None"
Pipeline version to run.
trace_id
Optional[str]
default:"None"
Trace ID for observability.
A :class:RunHandler for the background run.
terminate
Pipeline.terminate(run_id: str) -> dict[str, Any]
The pipeline run ID to terminate.
Streaming
stream
Pipeline.stream(inputs: dict[str, Any], stream_all_outputs: bool = False, session_id: Optional[str] = None, node_input_overrides: Optional[dict[str, Any]] = None, version: Optional[str] = None) -> Generator[StreamChunk, None, None]
Dictionary of input values for the pipeline.
Whether to stream all outputs as they arrive.
session_id
Optional[str]
default:"None"
Optional session ID for run grouping and tracing. Groups multiple runs under the same session for analytics and observability. Note: this does not provide conversational memory to LLM nodes — each run’s LLM nodes only see the current inputs.
node_input_overrides
Optional[dict[str, Any]]
default:"None"
Optional per-node input overrides.
version
Optional[str]
default:"None"
Pipeline version to run.
returns
Generator[StreamChunk, None, None]
A generator of :class:StreamChunk objects.
Sharing & publishing
share
Pipeline.share(user_id: Optional[str] = None, org_id: Optional[str] = None, role: "Literal[viewer, editor]" = viewer, name: Optional[str] = None) -> dict[str, Any]
user_id
Optional[str]
default:"None"
User ID to share with.
org_id
Optional[str]
default:"None"
Organization ID to share with.
role
Literal[viewer, editor]
default:"'viewer'"
Permission role ("viewer" or "editor").
name
Optional[str]
default:"None"
Display name for the shared entity.
unshare
Pipeline.unshare(user_id: Optional[str] = None, org_id: Optional[str] = None) -> dict[str, Any]
user_id
Optional[str]
default:"None"
User ID to unshare.
org_id
Optional[str]
default:"None"
Organization ID to unshare.
publish
Pipeline.publish(title: str, description: str = '', tags: Optional[list[str]] = None, internal: bool = False) -> dict[str, Any]
Marketplace listing title.
Marketplace listing description.
tags
Optional[list[str]]
default:"None"
Optional list of tags.
If True, publish as internal-only.
\{"status": "success", "id": "\<marketplace_object_id>"\}
unpublish
Pipeline.unpublish(marketplace_id: str) -> dict[str, Any]
The marketplace listing ID to remove.
move_to_folder
Pipeline.move_to_folder(folder_id: str) -> dict[str, Any]
Versioning
revert
Pipeline.revert(version: str) -> dict[str, Any]
The local `Pipeline` object is **not** updated after this call.
Call `Pipeline.fetch(id=...)` to get the reverted state.
Semantic version string, e.g. "1.2.3".
Serialization
to_dict
Pipeline.to_dict() -> dict
from_json
Pipeline.from_json(data: dict) -> Pipeline
Pipeline.serialize_inputs(inputs: dict[str, Any]) -> dict[str, Any]
Types
Configuration objects, response shapes, and enums used by the methods above.
BumpLevel
Members
PATCH = "patch"MINOR = "minor"MAJOR = "major"
RunResult
Response from :meth:Pipeline.run / :meth:Pipeline.arun.
Fields
RunStatus
Response from :meth:Pipeline.run_status / :meth:RunHandler.run_status.
Fields
status
Literal[in_progress, completed, failed]
required
StartResult
Response from :meth:Pipeline.start background run initiation.
Fields
