SyGra 2.0.0 introduces Studio, an interactive environment that turns synthetic data generation into a transparent, visual craft. Instead of juggling YAML files and terminals, you compose flows directly on the canvas, preview datasets before committing, tune prompts with inline variable hints, and watch executions stream live—all from a single pane. Under the hood it’s the same platform, so everything you do visually generates the corresponding SyGra compatible graph config and task executor scripts.
.executions/.Let’s walk through this experience step by step.
Open Studio, click Create Flow, and Start/End nodes appear automatically. Before adding anything else:
repo_id, split, or file path, then click Preview to fetch sample rows.{prompt}, {genre}), so you know exactly what can be referenced inside prompts and processors.Once validated, Studio keeps the configuration in sync and pipes those variables throughout the flow—no manual wiring or guesswork.
Drag the blocks you need from the palette. For a story-generation pipeline:
gpt-4o-mini), write the prompt, and store the result in story_body.{story_body} inside the prompt, and output to story_summary.Studio’s detail panel keeps everything in context—model parameters, prompt editor, tool configuration, pre/post-process code, and even multi-LLM settings if you want parallel generations. Typing { inside a prompt surfaces every available state variable instantly.
Open the Code Panel to inspect the exact YAML/JSON Studio is generating. This is the same artifact written to tasks/examples/, so what you see is what gets committed.
When you’re ready to execute:
.executions/runs/*.json.After the run, download outputs, compare against prior executions, get metadata of latency and usage details.
SyGra Studio can also execute existing workflow in the tasks. For example, in the tasks/examples/glaive_code_assistant/ workflow — it ingests the glaiveai/glaive-code-assistant-v2 dataset, drafts answers, critiques them, and loops until the critique returns “NO MORE FEEDBACK.”
Inside Studio you’ll notice:
generate_answer and critique_answer) linked by a conditional edge that either routes back for more revisions or exits to END when the critique is satisfied.git clone https://github.com/ServiceNow/SyGra.git
cd SyGra && make studio
tasks/examples/glaive_code_assistant/graph_config.yamlSyGra Studio turns synthetic data workflows into a visual, user friendly experience. Configure once, build with confidence, run with full observability, generate the data without ever leaving the canvas.