← Back to the Software and tools tiles
Evaluation, testing and observability
10 records Generated from this page's data. Click a name to open its panel on the Software and tools page. Records resting on one source are tagged "one source".| Record | What it is | Place | Links | Confidence |
|---|---|---|---|---|
| AtlaCompanies | Trains AI models whose job is to check and score the answers of other AI models.AI evaluation | London | atla-ai.com/jobs/research-engineer · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. · ycombinator.com/companies/atla | one source |
| ComposalCompanies | Sends swarms of AI agents to use a product as its customers would, finding bugs before release. | London (—) | ycombinator.com/companies/composal · ycombinator.com/companies/industry/Developer%20T | one source |
| DiffblueCompanies | Makes an AI agent that writes unit tests for Java code automatically, so developers do not have to write the tests by hand.software testingProfile
Round 6 deep profile (30 Oct 2026), as the researchers wrote it; sources are listed in the record panel. | Oxford (operating base per releases); registered office London (EC4A 3TW) | Companies House · businesswire.com/news/home/20250331532523/en/Dif · businesswire.com/news/home/20251104720918/en · joplinglobe.com/region/national_business/new-dif · find-and-update.company-information.service.gov. | verified |
| ETIQ AICompanies | Testing tool that helps data scientists find and fix bias in their machine-learning models. | digicatapult.org.uk/about/press-releases/post/10 · etiq.ai/posts/etiq-ai-announce-eu900-000-seed-fu | one source | |
| Evidently AICompanies | Open-source tool for monitoring and evaluating machine-learning and LLM systems in production.open-source ML monitoringKey facts Team size 2 (YC profile, 2026-10); YC Summer 2021 batch; GitHub evidentlyai | London | ycombinator.com/companies/evidently-ai | one source |
| HumanloopCompanies | Made software for managing prompts and for testing and monitoring apps built on large language models; its team joined Anthropic in 2025 and the product closed. wound_downLLM evaluation | London (operating base as reported); registered office Cambridge (CB1 2LA) | Companies House · techcrunch.com/2025/08/13/anthropic-nabs-humanlo · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. | one source |
| LaminarCompanies | Open-source tracing and debugging platform that records every LLM call and tool use by AI agents.AI agent observabilityKey facts Team size 6 (YC profile, 2026-10); users named: Browser Use, OpenHands, Rye.com, Alai (2026-03); default observability tool in Browser Use docs (2026-03) | London | tech.eu/2026/03/17/agent-debugging-startup-lamin · ycombinator.com/companies/laminar | one source |
| ManaMindCompanies | Autonomous AI agents that play video games to find bugs and write reports while the game is being built.AI agents for game testingKey facts Design partners: Included Games, Crazy Labs (2026-04) | London | tech.eu/2026/04/30/manamind-secures-15m-to-devel | one source |
| MeticulousCompanies | Records real user sessions and replays them against each code change to catch interface bugs, including in AI-written code.automated UI testingKey facts Team size 15 (YC profile, 2026-10); YC Summer 2021 batch | London | ycombinator.com/companies/meticulous | one source |
| Unlikely AICompanies · also on Labs | Builds its own AI for decisions in regulated fields such as financial services, accounting and insurance, where answers must be explainable and correct.Profile
Labs Round 7 deep profile (30 Oct 2026), as the researchers wrote it; confidence: single_source (all from company pages; seed round already verified in the deal file). Sources Models and products (1)
| London | Companies House · unlikely.ai/ · uktech.news/ai/unlikely-ai-funding-20220915 · find-and-update.company-information.service.gov. · find-and-update.company-information.service.gov. | verified |