Acrux Core sits between your application and every LLM provider, so you can ship, change, and measure AI features without redeploying code just to move a prompt.
It brings together the pieces teams normally stitch from separate tools:
It's multi-tenant, team-scoped, open soufor TypeScript/Node (@acruxcoreai/sdk onnpm) and Python (acruxcore on PyPI). Free while in beta.
Prompts
Versioned, templated message sets. Move the production alias to a new version and the next call picks it up โ no redeploy, no release, no code change.
Gateway
One OpenAI-compatible endpoint in front of every provider. Bring your own keys, issue scoped virtual keys with spend budgets, and get cost and cache-hit info on every response.
Tracing
Every gateway call is recorded as a trace with spans for model, tokens, latency, and cost โ grouped into sessions, with human feedback attachable to any span.
Tools
Callable functions versioned and promoted like prompts, attached directly to a prompt so one render call returns both, with every call recorded as an analytics span.
Evaluation
Build datasets from real traces and feedback, sweep them across prompt versions and models as experiments, and read per-cell results before promoting a prompt to production.
Most LLM tools own one link in the chain โ a prompt manager, a gateway, a tracer, or an eval tool โ and teams stitch three or four of them together with glue code. Acrux Core keeps one lineage instead: a prompt version flows through the gateway as a request, becomes a trace with tool spans, gets priced, and folds into an evaluation dataset โ all in one data model, so a regression can be traced back to the exact prompt version and model that caused it without cross-referencing four dashboards.
Because the pieces are already wired together. You don't run a prompt CMS, a separate LiteLLM-style proxy, a separate tracing tool, and a separate eval harness and reconcile them yourself โ one gateway call is already versioned, traced, and priced, and that trace can become an evaluation dataset with no export/import step. It's also open source, self-hostable, and ships official SDKs (TypeScript and Python) rather than locking you into a dashboard-only workflow.
Engineering teams shipping LLM features in production โ usually a backend or platform engineer who owns the AI integration for a small-to-mid-size product team. They need to change a prompt without a redeploy, see what a call actually cost and how long it took, and prove a new prompt version is better before promoting it โ without adopting a full MLOps stack to get there
TypeScript end to end โ an Express API and a BullMQ-based worker on Node.js, a React dashboard, and PostgreSQL (with JSONB for flexible payloads like prompt content and trace data) via Prisma as the ORM. Prompt templating uses a Jinja2-style engine (nunjucks). The public SDKs are TypeScript/Node (@acruxcoreai/sdk) and async Python (acruxcore). The docs site runs on Docusaurus.
I've spent the last three years building chatbots for more than 30 clients. Across all those projects, I kept running into the same problem with the tools already on the market: they were either bloated with features I never needed, or missing the ones I actually did. I'd end up paying for complexity I didn't want, while still gluing together separate tools for prompt versioning, request routing, tracing, and evaluation myself. After enough of that, I decided to build the platform I actually needed โ one place to version prompts, route calls to any model provider, trace every request, and evaluate whether a change actually made things better. Acrux Core is that platform, built from real production experience rather than a feature checklist.
We have collected here some useful links to help you find out if Acrux Core is good.
Check the traffic stats of Acrux Core on SimilarWeb. The key metrics to look for are: monthly visits, average visit duration, pages per visit, and traffic by country. Moreoever, check the traffic sources. For example "Direct" traffic is a good sign.
Check the "Domain Rating" of Acrux Core on Ahrefs. The domain rating is a measure of the strength of a website's backlink profile on a scale from 0 to 100. It shows the strength of Acrux Core's backlink profile compared to the other websites. In most cases a domain rating of 60+ is considered good and 70+ is considered very good.
Check the "Domain Authority" of Acrux Core on MOZ. A website's domain authority (DA) is a search engine ranking score that predicts how well a website will rank on search engine result pages (SERPs). It is based on a 100-point logarithmic scale, with higher scores corresponding to a greater likelihood of ranking. This is another useful metric to check if a website is good.
The latest comments about Acrux Core on Reddit. This can help you find out how popualr the product is and what people think about it.
Do you know an article comparing Acrux Core to other products?
Suggest a link to a post with product alternatives.
Is Acrux Core good? This is an informative page that will help you find out. Moreover, you can review and discuss Acrux Core here. The primary details have been verified within the last quarter. So they could be considered up to date. If you think we are missing something, please use the means on this page to comment or suggest changes. All reviews and comments are highly encouranged and appreciated as they help everyone in the community to make an informed choice. Please always be kind and objective when evaluating a product and sharing your opinion.