RESEARCH & ARCHITECTURE

BetterQuery: a practical architecture for intent-aware query enhancement

We are documenting the product as a production system, not a prototype. This section keeps research artifacts, technical papers, and shareable references discoverable for your team, investors, and integrator users.

Primary reference and contact: research@cdli.ai

System-first architecture

Skill-aware routing + semantic cache

Requests are intentionally routed before generation to preserve behavior and reduce bad outputs.

Defensive execution

Quality gates and controlled fallback

Fallback and degraded modes are explicit so usage remains predictable during model drift.

Measurable iteration

Admin metrics + session feedback

Top skills, fallback rates, cache hit behavior, latency, and user quality signals are persisted.

Research collection

All papers are in copy-ready LaTeX and aligned with current BetterQuery behavior, including MCP and extension distribution, routing strategy, and analytics evolution.

Academic-ready

Core technical paper

Intention Bridge Architecture

Full architecture and routing model for intent-first enhancement. Includes hybrid routing, caching, fallback, observability, and an evaluation protocol.

Author: CDLI Lab

1-page concise version

Intention Bridge — Executive Brief

Short version for product reviews, internal sharing, and proposal decks. Same assumptions, lighter format.

Author: CDLI Lab

Reference draft

Prism + OpenAI LaTeX Editor

Research writing sample used to validate math-heavy output formatting and citation quality with LaTeX examples.

Author: CDLI Lab

Per-paper citation blocks

Use these per-document references for proposals, white-paper attachments, or PRD appendices.

betterquery_intention_bridge_arch_2026

Intention Bridge Architecture

APA

CDLI Lab. (2026). BetterQuery Intention Bridge: Architectural foundations, operational guarantees, and validation plan. BetterQuery Research Technical Paper. https://betterquery.app/research/betterquery-intention-bridge.tex

BibTeX

@techreport{betterquery_intention_bridge_arch_2026,
  title = {BetterQuery Intention Bridge: Architectural Foundations, Operational Guarantees, and Validation Plan},
  author = {{CDLI Lab}},
  institution = {CDLI},
  year = {2026},
  note = {BetterQuery Research Technical Paper},
  url = {https://betterquery.app/research/betterquery-intention-bridge.tex}
}

betterquery_intention_bridge_brief_2026

Intention Bridge — Executive Brief

APA

CDLI Lab. (2026). BetterQuery Intention Bridge: Technical summary. BetterQuery Research brief. https://betterquery.app/research/betterquery-intention-bridge-short.tex

BibTeX

@techreport{betterquery_intention_bridge_brief_2026,
  title = {BetterQuery Intention Bridge: Technical Summary},
  author = {{CDLI Lab}},
  institution = {CDLI},
  year = {2026},
  note = {BetterQuery Research Brief},
  url = {https://betterquery.app/research/betterquery-intention-bridge-short.tex}
}

prism_openai_latex_editor_sample_2026

Prism + OpenAI LaTeX Editor

APA

CDLI Lab. (2026). Prism + OpenAI LaTeX editor sample: Citation quality and formatted output examples. CDLI Lab reference artifact. https://betterquery.app/research/prism-openai-latex-editor.tex

BibTeX

@misc{prism_openai_latex_editor_sample_2026,
  title = {Prism + OpenAI LaTeX Editor Sample},
  author = {{CDLI Lab}},
  howpublished = {Research artifact},
  year = {2026},
  note = {Reference sample used for citation and formatting validation},
  url = {https://betterquery.app/research/prism-openai-latex-editor.tex}
}

API entrypoint

The production API is the primary integration path and includes observability context in every response.

POST https://betterquery.app/api/v1/enhance

{
  "query": "Explain this codebase to a new engineer",
  "options": {
    "mode": "transparent"
  }
}

MCP command-line bridge

MCP is the fastest path for CLI/agent workflows. One install, one key, one intention-aware tool surface.

npx -y @cdli-ai/betterquery-mcp
BETTERQUERY_API_KEY=bq_live_...
View MCP package on npm →

Next research priorities

  • Multi-agent consensus scoring for routing and quality arbitration.
  • Offline benchmark release with public methodology and date-based splits.
  • Public changelog of routing policy updates and threshold history.
  • Research notes on cache economics and fallback cost under long-tail traffic.

Citation-ready references

Copy and paste these reference blocks when filing proposals, reports, or research docs.

APA

CDLI Lab. (2026). BetterQuery Intention Bridge: Architectural foundations, operational guarantees, and validation plan. BetterQuery Research Technical Paper.

BibTeX

@techreport{betterquery_intention_bridge_2026,
  title  = {BetterQuery Intention Bridge: Architectural Foundations, Operational Guarantees, and Validation Plan},
  author = {CDLI Lab},
  institution = {CDLI},
  year = {2026},
  note = {BetterQuery Research},
  url = {https://betterquery.app/research}
}

Want to read quickly? Start with the executive brief, then move into the architecture paper.