Prompts
LLM labeling tool inside Label Studio Enterprise: a team writes and versions prompts against a chosen model (OpenAI, Gemini, Claude, Azure OpenAI or its own endpoint), scores each version against ground-truth annotations for accuracy and inference cost, then pre-labels text and image tasks for human review.
What it does
- Data labelling
- Evaluation and observability
Features 4
- Pre-labeling with LLMsRapidly pre-label text and images for large datasets using off-the-shelf models or ground truth.
- Model comparisonPrompts versioning lets teams test models and measure accuracy scores and inference costs.
- LLM-as-a-judgeUse LLM-as-a-judge to detect toxicity when evaluating model outputs.
- Self-improving promptsLabel Studio can enhance prompts by referencing the user's data and chosen LLM.
Sources
- Pricing
-
The Prompts page routes buyers to Starter Cloud or to sales. Pricing page: Starter "$99 per month", "$49 per month for each additional user", free trial; Enterprise "Custom pricing". No separate price for Prompts is published, and model usage runs on the customer's own LLM API key (docs: "All you need to get started is an LLM deployment API key and a project."). The free access offered at launch was "for a limited time" and is a trial, not a free plan.
- Description
-
Company disclosure · 11 Oct 2026
"Save and edit multiple versions of prompts so you can refine and test across different models"; "Evaluate overall accuracy, outputs, F1 score, and inference costs"; supported providers OpenAI, Google Gemini, Vertex AI, Anthropic Claude, Azure OpenAI and "Custom Model Endpoints"; text and image labeling.
- What it does
-
Official documentation · 11 Oct 2026
"Automatically classify and annotate text and images with AI"; "Compare models for costs and quality ... or use LLM-as-a-judge to detect toxicity."
- Sold within
-
Company disclosure · 11 Oct 2026
"If you're on a Starter Cloud or Enterprise SaaS plan, you can start using Prompts today"; the current product page adds "talk to sales about adding Prompts to your plan", so on an Enterprise contract it may be enabled on request rather than by default.
- Name
-
Official documentation · 11 Oct 2026
"With Prompts in Label Studio Enterprise, we've been able to bootstrap our labeling performance to near-human accuracy"; "Ready to put Prompts to work?"
- URL
-
Official documentation · 11 Oct 2026
Prompts' own page, listed in the Label Studio Enterprise sub-menu (Overview, Interfaces, Automation, Quality Assurance, Data Security, Prompts).
- Deployment
-
Official documentation · 11 Oct 2026
Docs: "For information on installing Prompts for on-prem environments, see Install Prompts"; GA post: generally available in Label Studio SaaS.
- Interfaces
-
Official documentation · 11 Oct 2026
"Go to the Prompts page and create a new Prompt" inside the Label Studio web app.
- Modalities
-
Company disclosure · 11 Oct 2026
"Text Data: classification, named entity recognition, sentiment analysis, summarization"; "Image Data: captioning, classification"; outputs are predictions converted to annotations.
- Intended for
-
Company disclosure · 11 Oct 2026
Written for data scientists, machine learning teams and subject matter experts on Label Studio SaaS plans.
Also from HumanSignal
-
Label Studio Enterprise Platform
A paid data-labeling, evaluation and quality-control platform for AI teams, built on the open-source Label Studio tool.
-
HumanSignal Services Data service
Managed data-collection and annotation service for training data that cannot be scraped, licensed or crowdsourced: HumanSignal designs the collection protocol, runs the operation and delivers a verified dataset, with evals, RLHF and red-teaming work.
Open HumanSignal in the directory
Something wrong here? Send a correction — quote this product id: humansignal-prompts.