# Next Moca

> Next Moca builds Needlepath (the context relevance engine: selects what matters from
> agent state before every model call, +12.9 pp better answers and 52.9%
> fewer input tokens delivered than sending full context on a public benchmark, measured at
> the np-2026-08-r4 operating point, counting the 24.0% of calls
> that received the complete document) and the Agent
> Control Plane (the turnkey white-label platform agent providers use to ship agentic AI
> in their customers' VPCs, on any model).

Machine-readable mirror: every page listed below has a markdown variant at the .md URL.
Single-file version: https://www.nextmoca.com/llms-full.txt

## Products

- [Needlepath: The Context Relevance Engine](https://www.nextmoca.com/index.md): selection layer, not a compressor; +12.9 pp better answers, 52.9% fewer input tokens delivered (counting the 24.0% of calls that received the complete document), 25.0 ms mean per selection decision, cheaper than full context all in ($38.44 vs $69.02 on the same 2,600 answers); stands aside honestly when full context is the right packet
- [RULER results in full](https://www.nextmoca.com/research/ruler): the run behind those figures, with the refusal rate, the delivered-token counts, the cost per correct answer, and the selector latency distribution
- [Agent Control Plane](https://www.nextmoca.com/products/control-plane.md): white-label agent orchestration in the customer's VPC; sovereignty, model neutrality, compounding knowledge; founding customer Aurivio AgentBase live in production

## Research and foundations

- [Foundations index](https://www.nextmoca.com/foundations.md): the work behind the product
- [RULER results in full](https://www.nextmoca.com/research/ruler): Needlepath against full context on the RULER public suite, 2,600 questions, one run: accuracy, delivered tokens, refusal rate, cost per correct answer, selector latency
- [FinanceBench results](https://www.nextmoca.com/research/financebench): the fit-limits page, Needlepath against full context on the public FinanceBench test set by Patronus AI, 70 questions: where a smaller packet could not be shown safe, the whole filing went through
- [Aethon](https://www.nextmoca.com/foundations/aethon.md): instant agent instantiation through reference-based replication
- [Tool Forge](https://www.nextmoca.com/foundations/tool-forge.md): validation-carrying toolchain for governed agentic execution
- [Public benchmark harness](https://github.com/nextmoca/context-selection-bench): reproducible matched-protocol benchmark for context-selection methods
- [Agent Definition Language (ADL)](https://github.com/nextmoca/adl): open standard for defining AI agents

## Blog

- [Blog index](https://www.nextmoca.com/blog.md): posts on agent infrastructure, ADL, ADLC, benchmarks, and weekly AI wraps (each post: append .md to its URL where available)

## Legal

- [Terms of Service](https://www.nextmoca.com/terms.md)
- [Privacy Policy](https://www.nextmoca.com/privacy.md)

## Company

Founded by Kiran Kashalkar (Co-CEO, ex-Oracle OCI, ex-Brightcove, Babson Butler
Launchpad) and Swanand Rao (Co-CEO, ex-Oracle, M9 exit, architect of the Federation
Model). Backed and built with NVIDIA Inception, Google for Startups, AWS Activate,
Babson Butler Launchpad, and ND Labs. Boston · Palo Alto.

Contact: kiran@nextmoca.com · swanand@nextmoca.com · https://www.nextmoca.com/#cta