A

Ashr Evals

Platform

by Ashr

Testing platform for AI agents: generates synthetic scenarios (text, voice and image turns), runs a team's agent against each through a Python or TypeScript SDK, grades prompts and tool calls and triages failures in a workspace, and traces the agent in production; Developer at $39 a month and Startup at $500 a month with metered overage, and a contact-sales Enterprise plan.

Find alternatives to Ashr Evals

What it does

• Evaluation and observability

Sources

Pricing

Pricing page · 21 Oct 2026

Plans section "Scale agent testing by usage.": Developer $39 a month (hard cap), Startup $500 a month (metered overage), Enterprise custom with a Contact sales button; no free plan is listed.

Description

Official documentation · 21 Oct 2026

"Test agents with real environments. Run evals, catch regressions, fix failures — before your users find them."; plan features "Synthetic scenario generation", "Prompt and tool-call grading", "Failure triage workspace".

What it does

Official documentation · 21 Oct 2026

"Test AI agents against real environments and catch regressions before your users do."

Sold within

Pricing page · 21 Oct 2026

"Developer and Startup are the same Ashr product with different included limits. Enterprise stays high-touch with a contact sales flow." Each paid plan has its own price and button.

Name

Official documentation · 21 Oct 2026

Page title and heading "Ashr Evals"; hero "Evals for Agents That Actually Work".

URL

Official documentation · 21 Oct 2026

Own page, linked from the Ashr menu as "Evals PLATFORM".

Deployment

Official documentation · 21 Oct 2026

The pages read state no deployment model beyond the vendor-run dashboard (lab.ashr.io); saas is recorded as the vendor-operated reading.

Interfaces

Official documentation · 21 Oct 2026

Menu: "Python & TypeScript SDK reference, quickstarts, and examples."; the SDK docs say "Traces render in the Observability panel of the dashboard."

Modalities

Official documentation · 21 Oct 2026

Scenario turns on the page are labelled Audio, Chat, Image and File ("Customer sends photo of item"); plans meter "Voice minutes"; results read "graded: 13 / 15 passed".

Intended for

Official documentation · 21 Oct 2026

Developer plan: "For teams validating their first production agent."

Also from Ashr

• Ashr Platform

  Agents that turn a company's traces, documents and logs into a training set, post-train open-weight models, grade each candidate in the customer's own environment and serve it through a downloadable set of weights or an OpenAI- and Anthropic-compatible endpoint, through the Manifold workbench; engineers deploy alongside the agents and no prices are published.

Open Ashr in the directory

Something wrong here? Send a correction — quote this product id: ashr-evals.