Skip to main content
DocsAPI Open PlatformOfficial SDK
Open platform · 12 languages

Braidrun SDK

Use 13 type-safe clients covering 12 languages to call workflow, execution, artifact, approval, and webhook APIs.

Braidrun SDK is the official collection of type-safe clients for the open platform API. It wraps workflow, execution, artifact, human approval, and webhook endpoints in familiar methods and data models for each language, so you do not have to write HTTP requests and response types by hand.

Official source repositorygithub.com/Braidrun/braidrun-sdkGenerated from the same OpenAPI 3.0.3 specification, all clients share consistent API 1.0 interfaces and models.
View repository

Supported clients

The repository provides 13 client implementations covering 12 languages. Each directory is a standalone client with package metadata, API models, tests, and complete endpoint documentation for that language.

API coverage

Every language client covers the same public capabilities, with method names adapted to the conventions of the language.

CapabilityWhat you can do
WorkflowList and read workflows, and start an execution asynchronously.
ExecuteQuery status, history, and final results, or cancel an execution that is still running.
ArtifactRead files and artifact metadata produced by an execution.
Manual ApprovalRead pending approvals and submit an approve or reject decision.
WebhookRegister and manage webhooks, view statistics, and trigger workflows through webhooks.

Quick Start

  1. Create a key under “Account Settings → API Keys” and grant only the scopes required for your calls.
  2. Choose the client that matches your technology stack and open the README in its directory.
  3. Install the dependencies from source and configure the API Key through an environment variable.
  4. Create an API client, then call the workflow list first to verify the connection and permissions.
terminalbash
python -m pip install "git+https://github.com/Braidrun/braidrun-sdk.git#subdirectory=python"
list_workflows.pypython
import os
import braidrun_sdk

configuration = braidrun_sdk.Configuration(
    host="https://braidrun.com",
    access_token=os.environ["BRAIDRUN_API_KEY"],
)

with braidrun_sdk.ApiClient(configuration) as client:
    workflows = braidrun_sdk.WorkflowsApi(client)
    result = workflows.list_workflows()
    print(result)
Currently distributed as source

The SDK is currently distributed through its GitHub source repository, and availability in language-specific package managers may vary. Refer to the README in the relevant directory; do not assume that every client has been published to a public package repository.

Authentication and permissions

Load the API Key beginning with dyk_ from an environment variable or secret management service, and use a Bearer header. Never write a real key into source code, logs, or browser frontend code. Each endpoint still checks the required scope; a type-safe client does not grant permissions automatically.

Webhook management uses different authentication

The SDK also generates webhook management APIs, but management endpoints under /api/webhooks use a browser session or an administrator service account and do not accept standard fine-grained open platform API Keys. Webhook trigger endpoints can still use an open platform key with the WEBHOOK_TRIGGER scope.

When the API returns 429, honor Retry-After and use X-RateLimit-Limit and X-RateLimit-Window to adjust your retry cadence.

SDK or workflow engine?

ProjectSuitable For The SceneRuns in
braidrun-sdkCall the Braidrun cloud platform API from your application with language-specific types and models.Your application + Braidrun cloud platform
braidrun-workflowRun YAML workflows in your own environment or embed the runtime in a Kotlin/JVM project.Local, server, or CI environments you manage

See the full overview of both projects