- Platform
- Preview
- Labs
Platform is the composable core of Parable. Everything here has been seen across enough customers, long enough, that we’re confident it behaves the same way for yours. We move things into Platform deliberately — a great composable platform takes patience to build, and we’d rather be right than fast at this layer. If something’s in Platform, it’s production, stable, and ours to keep steady.
How to read Platform results: this is the bedrock.
How to read Platform results: this is the bedrock.
New
Syncs now pick up only what's new since the last run
Syncs now pick up only what's new since the last run
Direct impact
- Parable’s newest ingestion pipelines now remember where they left off and sync only the fresh data on each run.
- Fewer duplicate pulls, faster sync times, and less load on your connected vendors — especially for high-volume sources like email, chat, and documents.
- Happens automatically; no setup required on your side.
Built-in support for file and image uploads across the platform
Built-in support for file and image uploads across the platform
- The platform now speaks native file, image, and logo uploads — any future Parable feature that needs to accept an image or document from you (for example, customising how your workspace is branded, or attaching supporting files) shares one reliable upload path with previews, validation, and CDN-backed delivery.
- You won’t see this directly this week; it’s the foundation for features that land over the coming cycles.
Search and sort across tenants in the admin app
Search and sort across tenants in the admin app
- If you manage multiple workspaces from Parable’s admin surface, you can now search by name and order the list by creation date.
- Small quality-of-life improvement for teams that juggle many tenants day to day.
Improved
Faster, leaner in-app experience
Faster, leaner in-app experience
- Behind the scenes, the web app now calls Parable’s backend directly instead of hopping through an internal proxy layer.
- Result: fewer moving parts between your click and the answer, and a simpler security path for the engineers maintaining it.
Steadier background infrastructure
Steadier background infrastructure
- Our pipeline workers are now actively watched for memory, CPU, and out-of-memory events, with alerts that wake someone up before a silent failure becomes a customer-visible one.
- This follows an incident last week where a worker node was killed by the operating system — the VM was already upsized; this week adds the early-warning system to prevent a repeat.
Fixed
Slug-based lookups now work end-to-end
Slug-based lookups now work end-to-end
- A codegen bug had been producing broken database lookups for any field keyed by a human-readable slug (for example, “look up this vendor by name”). Fixed, along with a handful of related issues around optional relations and null filters.
- You’d only have felt this if you were on a very recent build; it’s resolved now.
How we tier releases: Labs · Preview · Platform. The world — and the AI beneath every intelligent product — is changing faster than ever. We ship you the latest and greatest. Occasionally your results will shift as models and methods refine; rarely enough to flip a decision, always with a note from us when the change is meaningful.