q2
Ten releases in six days, methodically porting Quarto's surface into a Rust binary.
A side-by-side editorial comparison of evaluate and WorkOS — release velocity, themes, recent moves, and the top alternatives to consider.
| Feature | evaluate | WorkOS |
|---|---|---|
| Sector | Infra & APIs | Infra & APIs |
| Velocity score | 0.0 | 8.8 |
| Sparks · 30d | 0 | 3 |
| Top themes | r-lib, code-evaluation, knitr, graphics-capture | agent identity, authentication, token custody, developer experience |
| Last editorial update | 1h ago | 4h ago |
| Website | Visit → | — |
The engine under every knitted R document reached 1.0 by making evaluation behave like the console.
evaluate captures the output, plots, messages and conditions produced by running R code, and it is the layer knitr and R Markdown sit on. The 1.0.0 release changed core semantics — multi-expression input now stops at the first error — and gave results a real class. Since then, releases have been graphics-focused: ragg-based capture when available, grid plot fixes, and a patch for ggplot2 4.0.0.
WorkOS is building the identity layer for agents, then shipping an agent of its own on top of it.
WorkOS has spent the last six weeks assembling the pieces an agent needs to act on a user's behalf: Agent Registration gives agents their own AuthKit identity via the auth.md protocol, the Pipes Token Proxy calls third-party APIs without the application ever touching a user's access token, and Pipes now covers API key providers as well as OAuth. A Management MCP Server and one-click plugins for Claude and ChatGPT put the platform itself inside those agents. Atlas, an AI coworker in Slack, sits on top as an end-user product rather than a developer primitive.
evaluate captures the output, plots, messages and conditions produced by running R code, and it is the layer knitr and R Markdown sit on. The 1.0.0 release changed core semantics — multi-expression input now stops at the first error — and gave results a real class. Since then, releases have been graphics-focused: ragg-based capture when available, grid plot fixes, and a patch for ggplot2 4.0.0.
The package is being tightened rather than expanded. 1.0.0 removed complexity it could not justify, deprecated include_timing, dropped a function testing for an object it never created, and aligned error behaviour with the console; 1.0.4 then moved plot capture toward a faster, more capable device. The recurring theme is reliability of output capture across environments, which is exactly what a document-rendering substrate needs and what local_reproducible_output() exists to enforce.
Expect continued fast-follow patches keyed to graphics-stack releases — ggplot2 and ragg in particular — rather than new API surface.
WorkOS has spent the last six weeks assembling the pieces an agent needs to act on a user's behalf: Agent Registration gives agents their own AuthKit identity via the auth.md protocol, the Pipes Token Proxy calls third-party APIs without the application ever touching a user's access token, and Pipes now covers API key providers as well as OAuth. A Management MCP Server and one-click plugins for Claude and ChatGPT put the platform itself inside those agents. Atlas, an AI coworker in Slack, sits on top as an end-user product rather than a developer primitive.
The sequence reads as a deliberate stack rather than scattered AI features — identity for the agent, custody of the credentials it uses, then an application that demonstrates both. Around it runs unglamorous platform work that suggests the same customers are being taken seriously in production: a local API for CI/CD with seeded data, signed webhooks and fault injection, dashboard read-only roles, Radar extended to the User Management APIs, and smaller onboarding fixes. The developer surface keeps widening by framework, with AuthKit for Astro.
Expect authorization and audit to follow registration — per-agent scopes, revocation and a record of what an agent did on whose behalf — since agent identity without those is only half of what the Pipes proxy position implies.
Other Infra & APIs products tracked by Sparkpulse, ranked by recent ship velocity. Each card links to a full editorial trajectory and lets you pivot into a head-to-head comparison with either evaluate or WorkOS.
Ten releases in six days, methodically porting Quarto's surface into a Rust binary.
The metadata cache under pak now speaks to authenticated and corporate repositories.
R's old dependency manager now runs on a vendored copy of its own successor.
A fake web server for tests that keeps closing the gap with the real thing.
Hit 1.0, relicensed to MIT, and has shipped almost nothing but dbplyr compatibility since.
carrier's 0.3.0 quietly rewrites how crated functions see each other — and breaks code that relied on the old rules.
See all evaluate alternatives → · See all WorkOS alternatives →
Latest ship moves from both products, interleaved chronologically. ⚡ = editorial spark.
They serve adjacent needs but don't currently overlap on shipped themes. WorkOS is currently shipping more aggressively (velocity 8.8 vs 0.0), with 3 editorial sparks in the last 30 days against 0. See the at-a-glance table above for a side-by-side breakdown of velocity, recent sparks, and editorial themes.
Sparkpulse doesn't pick a winner — we score release velocity, not feature parity. WorkOS is currently shipping more aggressively (velocity 8.8 vs 0.0), with 3 editorial sparks in the last 30 days against 0. For your specific use case, the alternatives sections above list other Infra & APIs products to evaluate alongside.
Top evaluate alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "evaluate alternatives" section above for the current picks, or visit /alternatives/evaluate for the full list with editorial commentary on each.
Top WorkOS alternatives in Infra & APIs are ranked by recent ship velocity. Browse the "WorkOS alternatives" section above for the current picks, or visit /alternatives/workos for the full list with editorial commentary on each.