Skip to content
TrustList
WB
Automated Testing Software

Wren by Seldon

AI agents that test a web application from plain-English YAML tests, run in parallel and record each step with evidence and reproduction steps; first 10 runs free, then $3 per completed run.

TrustList summary

Not yet independently verified. Dated from the Product Hunt feed entry of 9 November 2026 because the maker's page gives no launch date; the price is the maker's own. We will update this when it can be confirmed, and remove this note.

Test suites tend to rot: scripted browser steps break when a button moves, and nobody writes the case for the bug nobody expected. Wren, the software QA product from Seldon, replaces the scripts with tests written in plain sentences and runs them with AI agents that explore the application in parallel.

Tests in plain English, defined in YAML

A test is described in a sentence or two, kept in YAML, and run without specifying each browser action. Seldon contrasts this with Playwright scripts, where every click and field is spelled out in code. Many runs go at once, so a suite that would take a person an afternoon finishes while the team waits.

Agents guided by behavioural models

Seldon is an applied research lab that models how reasoning, emotion, experience and social influence shape decisions, and Wren draws on that work. The models guide what each agent tries next, which the maker says can surface test cases the team did not write. It names broken layouts, failed API calls and incorrect totals as the kinds of fault it looks for. Seldon's other products are Finch, for usability testing, and Rook, for national security and defence.

Write, run, find

Every run is recorded step by step, with evidence for each failure and the steps to reproduce it, so an engineer can see what the agent did before deciding whether it is a real defect. The maker suggests trying Wren alongside an existing suite rather than replacing it, and offers volume pricing and enterprise agreements for larger teams.

Pricing

The first 10 completed runs are free and shared across the workspace. After that each completed run costs $3. Model usage, browser execution, findings and recorded evidence are included, with no separate token charges and no seat fees, and runs that fail to finish because of an agent or platform fault are not charged. The price is the maker's own.

Key facts

Product: Wren, AI QA test agents Maker: Seldon (Blok Intelligence Inc.) Launched: Product Hunt feed, 9 November 2026 Pricing: first 10 completed runs free; then $3 per completed run; volume and enterprise on request Tests written as: plain English in YAML Compliance listed by maker: SOC 2, GDPR, HIPAA, ISO 27001 Website: seldon.com/products/wren

Categories & features

Request a demo or quote from Wren by Seldon

Protected by reCAPTCHA - Google Privacy Policy and Terms apply.

By sending, you agree we may share your request and contact details with the provider once you confirm your email.

Reviews

Write the first review of Wren by Seldon

Used it? Your experience helps other buyers decide.

Write a review

Questions & answers

No questions yet. Be the first to ask about Wren by Seldon.

Wren by Seldon alternatives

Similar listings buyers compare against this one.

London-built swarm of browser agents that walks through a web product the way users would, files each failure with a recording, logs and steps to reproduce, and hands it to a coding agent to fix and re-test.

Checks that code written by AI coding agents does what was asked, keeping specs versioned beside the code. From the team behind Qase; plans start at $49 a month, per its pricing page.

Turns browser end-to-end test runs into narrated product demos that are re-checked on a schedule, so a demo cannot drift from what shipped. Open-source CLI; hosted Team plan at $39 a month.

Evaluation and observability platform for WebMCP tools, testing them before release and monitoring agent use in production; free plan, Pro at $65.83 a month.

Continuous-integration check for web apps that replays critical user journeys on every GitHub pull request and blocks broken merges

Simulated, stateful copies of Gmail, Slack, Salesforce, Notion and QuickBooks where teams test AI agents before letting them touch real accounts, with worlds built from a description or from production traces.

Scriptable in-memory IMAP server for testing mail client software, with 50+ extensions and scripted faults