
- Monthly visits
- 2,890
- 2026-09
- Monthly growth +400.9%
Evaligo - Build AI workflows and evaluate language models
Evaligo provides a visual AI workflow builder with model integrations, web scraping, browser automation, data processing, testing, and API deployment. Its site also presents benchmarks and reinforcement-learning environments for training and evaluating LLM agents on public and held-out tasks. Users can compare workflow runs or benchmark results and choose whether to publish held-out evaluation results.
- AI Agents
- Workflow Automation
- LLM Evaluation
- Model Training
- AI workflow builder
- LLM agent benchmarks
- AI workflow automation




Capabilities
Visual workflow builder
Build AI workflows in a drag-and-drop editor and run them as workflows or API endpoints. Add prompt nodes, variables, conditional logic, data transformations, integrations, and execution schedules.
Workflow testing and comparison
Test workflows while building and inspect data at each node. Run agentic flow variants side by side to compare quality, cost, and latency, and review execution history and datasets.
Web data automation
Collect and process web data within workflows. The available actions include scraping pages, searching Google, browser interaction, HTTP requests, CSS or XPath extraction, and link extraction.
LLM agent benchmarks
Evaluate language models and agent scaffolds on task benchmarks, including SWE-Race and SWE-Pair. Public tasks can be run with a connected model; held-out tasks produce an official score.
Reinforcement-learning environments
Train models on task environments built from real work, with simulated tools and separate training and test splits. The environments use Harbor format and can be used with RL trainers.
Pricing
Evaligo offers a free plan and a paid monthly Pro subscription. Compare monthly credits, scheduling, concurrent API executions, and the ability to purchase additional credits when choosing a plan.
| Plan | Price | Included features and limits |
|---|---|---|
| Free | $0/month |
|
| Pro | $7/month |
|
Use cases
Build and deploy an AI workflow
When a task needs connected model calls and other automation steps, use the visual builder to assemble prompts, integrations, data operations, and conditions. Test the workflow and deploy it as an API endpoint.
Compare alternative AI workflow runs
When adjusting prompts, models, or flow settings, run variants side by side and inspect quality, cost, and latency. Use testing and execution history to review the resulting runs.
Evaluate an LLM agent on benchmark tasks
When comparing a model or agent scaffold, connect a model or endpoint and run public benchmark tasks. For an official held-out score, submit the scaffold for evaluation on the private split.
Train a model in an RL environment
When developing an LLM agent, run rollouts on the environment's training split using a compatible RL trainer. Evaligo describes the environment format as Harbor and reserves held-out tasks for evaluation.
Limitations
Free-plan AI nodes using Evaligo platform keys pause after the plan's credits are exhausted until the next billing cycle; using personal API keys avoids consuming those credits.
The Free plan provides one-time scheduled runs; recurring schedules are listed for Pro.
The public benchmark split does not provide the official score; Evaligo reserves that score for runs on its private held-out split.
The private evaluation results remain private unless the user chooses to publish them.
Team features are described as in development; the pricing page does not present a currently available team or enterprise plan.
Who it is for
People building AI workflows that combine models, web data, and integrations
Teams developing LLM agents that need task-based evaluation and training environments
Model developers comparing agent runs on public and held-out benchmark tasks
FAQ
What consumes credits on the Free plan?
AI model calls made with Evaligo platform API keys consume credits. The amount depends on the model and token usage, and the app shows credit usage in real time.
What happens when platform credits run out?
On Free, AI-powered nodes using platform keys pause until the next billing cycle. Users can upgrade to Pro or use their own API keys, which do not consume Evaligo credits.
Can I use my own model API keys?
Yes. The pricing page says OpenAI, Anthropic, and Google keys can be used on any plan, and calls made with those keys do not consume Evaligo credits.
Can Pro users purchase more credits?
Yes. Pro users can purchase additional credit packs, and the pricing FAQ says those packs do not expire.
What does a held-out benchmark score mean?
Evaligo runs the submitted scaffold on the private split to provide the official score. Results remain private unless the user chooses to publish them.
Is there a team or enterprise plan?
The pricing page says team features are in development and does not list a current team or enterprise plan. It directs users needing collaboration, SSO, or enterprise features to contact Evaligo.
Comments