Langbase is a serverless AI developer platform that enables developers to build, collaborate, and deploy AI agents and applications with composable AI infrastructure.
Langbase is a serverless AI developer platform that enables developers to build, collaborate, and deploy AI agents and applications with composable AI infrastructure. On Nagent, Langbase is exposed as a fully-configurable ai agents integration that any agent can call — 20 actions, and API key authentication. No code is required to wire Langbase into your workflow — connect it once via the External Integrations panel and reuse it across every agent you build.
Agent builders use Langbase to automate the kinds of tasks ai agents teams previously handled manually. Concrete examples — each one is a single agent step in Nagent — include:
Every action and trigger is paired with a structured input/output schema (visible in the sections below), so when you wire Langbase into Helix — our agentic agent builder — the editor knows exactly what each step expects and produces. Configure once, deploy anywhere across your Nagent agents.
Every operation an agent can call against Langbase, with input parameters and output schema. Drop these into any step of an agent built in Helix.
LANGBASE_APPEND_THREAD_MESSAGESTool to add new messages to an existing conversation thread. Use when continuing a chat session or adding context to a thread.
Input parameters
Array of one or more messages to append to the thread. Each message must have at minimum a 'role' field.
Unique identifier of the thread to append messages to.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_CHECK_HEALTHTool to check the health status of the Langbase API service. Use when you need to verify API availability.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_CHUNKER_SPLIT_CONTENTTool to split content into smaller chunks. Use when processing large text segments to fit downstream limits.
Input parameters
The input text content to be chunked.
Number of characters to overlap between consecutive chunks. Must be ≥256 and < chunk_size.
Desired size of each chunk in characters. Must be between 1024 and 30000. Default: 1024.
Mode for chunking. Use 'semantic' to preserve sentence boundaries; 'basic' for fixed-size splits.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_CREATE_THREADTool to create a new conversation thread. Use when starting a fresh chat session or grouping messages into a distinct thread.
Input parameters
Optional initial messages for the thread
Key-value pairs to store with the thread for organizational purposes
Custom thread identifier. If not provided, a unique ID will be generated.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_DELETE_THREADTool to delete a thread that is no longer needed to manage conversation history. Use when you need to permanently remove a thread by its ID.
Input parameters
The unique identifier of the thread to delete
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_DELETE_THREAD_MESSAGETool to delete a specific message from a conversation thread. Use when you need to remove a message from a thread by its ID.
Input parameters
The unique identifier of the thread containing the message
The unique identifier of the message to delete
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_DOCUMENT_LISTTool to list documents in a specific memory. Use when you need to fetch document metadata (and optionally vectors) from a memory after confirming its name. Supports pagination via limit and startAfter parameters.
Input parameters
Maximum number of documents to return
Name of the memory to list documents from
Pagination token to start after a specific document
Whether to include vector embeddings in each document
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_GET_PIPETool to retrieve details of a specific pipe by owner and name. Use when you need to fetch configuration and settings of a particular pipe.
Input parameters
The name of the pipe to retrieve
The login/username of the pipe owner
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_GET_THREADTool to retrieve details of a specific conversation thread. Use when you need the full thread details by its ID after confirming its existence.
Input parameters
The unique identifier of the thread to retrieve
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_LIST_MODELSTool to get available AI models supported by Langbase. Use to discover text and image generation models from various providers.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_LIST_THREAD_MESSAGESTool to list all messages in a conversation thread. Use after obtaining the thread ID to fetch its messages.
Input parameters
Number of messages to return per page.
Return messages created before this message ID.
Unique identifier of the conversation thread.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_LIST_TRACESTool to get execution traces for debugging and monitoring pipe runs. Use when you need to retrieve trace logs for a specific primitive.
Input parameters
Maximum number of traces to return
Order for trace results
Number of traces to skip for pagination
ID of the primitive (pipe) to get traces for
Name of the primitive (pipe) to get traces for
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_MEMORY_CREATETool to create a new memory. Use when storing a new memory record in Langbase after confirming memory details.
Input parameters
The name (identifier) of the memory.
Number of chunks to return. Default 10, range 1-100.
Maximum number of characters in a single chunk. Default 10000, min 1024, max 30000.
A short description of the memory.
Number of characters to overlap between chunks. Default 2048, min 256, must be less than chunk_size.
The embedding model to use. Supported: 'openai:text-embedding-3-large', 'cohere:embed-v4.0', 'cohere:embed-multilingual-v3.0', 'cohere:embed-multilingual-light-v3.0', 'google:text-embedding-004'. Defaults to 'openai:text-embedding-3-large'.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_MEMORY_DELETETool to delete a specific memory. Use when you need to permanently remove a stored memory by its name.
Input parameters
Name of the memory to delete
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_MEMORY_LISTTool to list all memory objects. Use when you need to fetch stored memories for context retrieval.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_PIPE_CREATETool to create a new pipe. Use after configuring pipe parameters. Returns an array of pipe objects, each including API key and URL.
Input parameters
Enforce the output to be in JSON format
Name of the pipe
Up to 4 sequences where the API will stop generating further tokens
Pipe LLM model in format 'provider:model_id'
If enabled, both prompt and completions will be stored
Nucleus sampling parameter between 0.0 and 1.0
Visibility status of the pipe
If enabled, the output will be streamed in real-time
Upsert pipe. If true, updates existing pipe with same name
If enabled, Langbase blocks flagged requests automatically
Maximum number of tokens in the response
Short description of the pipe
Sampling temperature between 0.0 and 2.0
Controls which (if any) tool is called by the model
Penalizes a word based on its occurrence in the input text
Penalizes a word based on how frequently it appears in the training data
If true, calls multiple tools in parallel
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_PIPE_LISTTool to list all pipes. Use after authentication to retrieve the complete list of pipes. Returns an array of pipe objects; callers must handle list iteration.
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_UPDATE_PIPETool to update an existing pipe's configuration on Langbase. Use when modifying model settings, parameters, prompts, tools, or memory. The pipe must already exist.
Input parameters
Enforce the output to be in JSON format
New name for the pipe
Up to 4 sequences where the API will stop generating further tokens
Pipe LLM model in format 'provider:model_id'
If enabled, both prompt and completions will be stored
List of tools to make available to the pipe
Nucleus sampling parameter between 0.0 and 1.0
List of memory objects to attach to the pipe
Visibility status of the pipe
If enabled, the output will be streamed in real-time
List of messages to configure the pipe's prompts
If enabled, Langbase blocks flagged requests automatically
Name of the pipe to update
Variables for the pipe, can be an array or object
Maximum number of tokens in the response
Short description of the pipe
Sampling temperature between 0.0 and 2.0
Controls which (if any) tool is called by the model. Can be a string like 'auto' or 'required', or a structured ToolChoice object
Penalizes a word based on its occurrence in the input text
If true, regenerates the API key for this pipe
Penalizes a word based on how frequently it appears in the training data
If true, calls multiple tools in parallel
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_UPDATE_THREADTool to update an existing thread's metadata. Use when you need to modify metadata fields for managing and organizing conversation threads.
Input parameters
Key-value pairs to store with the thread for organizational purposes. Each value must be a string with max length of 2048 characters.
The unique identifier of the thread to update
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
LANGBASE_UPDATE_THREAD_MESSAGETool to update an existing message in a conversation thread. Use when you need to modify the content or metadata of a specific message.
Input parameters
Updated text content of the message
Key-value pairs to store with the message. Each value must be a string with max length of 2048 characters.
Unique identifier of the thread containing the message
Unique identifier of the message to update
Output
Data from the action execution
Error if any occurred during the execution of the action
Whether or not the action execution was successful or not
No publicly available marketplace agent is found using this tool yet. There are 89 agents privately built on Nagent that already use Langbase.
Build on Nagent
Connect Langbase to any Nagent agent in minutes — no API key management, no boilerplate. Just configure and deploy.
The five questions agent builders ask before adopting a new integration.
Open the External Integrations panel inside Nagent (app.nagent.ai/externalIntegration), find Langbase, and click "Connect Now." You'll authenticate with an API key — Nagent handles credential storage and refresh automatically. Once connected, Langbase is available to any agent in your workspace.
No. Nagent provides no-code integration for every tool. Once Langbase is connected, you configure its 20 actions directly in the agent builder UI — no API calls, no boilerplate, no schema management.
Helix — Nagent's agentic agent builder — lets you drop Langbase steps into any workflow visually. Pick an action (e.g., one of those listed above), fill in the inputs (Helix knows the required vs. optional schema for each parameter), and connect it to upstream/downstream steps. Triggers run as the entry point of an agent, so when a Langbase event fires, the agent kicks off automatically.
Every Langbase action and trigger ships with a fully-typed schema — input parameters with name, type, required flag, and description, plus the output payload shape. The schemas are documented in the sections above. Helix uses these schemas to validate your configuration at build time and to type-check the data flowing between steps.
Yes. While Langbase ships with 20 pre-built ai agents actions, you can layer custom logic around them inside Helix — pre/post-processing steps, conditional branches, retries, or stitching Langbase together with other connected tools. For deeper customization, talk to our team about Nagent's Agentic AI Lab — forward-deployed engineers who build Langbase-based workflows tailored to your business.