Knowledge Base

Give every agent knowledge it can actually cite

Upload policies, rate cards, decks — even recorded training calls. Zoice chunks and embeds them for hybrid RAG retrieval, tags every chunk by language, and has agents append a "Sources:" footer to grounded answers.

  • PDF, Word, PowerPoint, Excel, CSV, JSON, Markdown — or paste raw text
  • Audio and video too: .mp3 .wav .ogg .m4a .mp4 .webm
  • Retrieval test bench — see the exact top-K chunks your agent receives

loan-products-kb

Attached to 3 agents

Hybrid RAG

Home-Loan-Policy.pdf

482 chunks · hi / en

Branch-Training-Call.mp3

Transcribed · 96 chunks

Rate-Card-Q2.xlsx

38 chunks · en

616

Total chunks

100%

Embedding coverage

412

Avg tokens / chunk

Rechunk recommended

2 files changed since last index

Chunk health updates automatically as your documents change.

How it works

From raw documents to grounded answers in four steps

Everything between upload and a cited answer is visible — chunk counts, retrieval rankings, and grounding checks included.

01

Upload anything

Drop in .pdf, .txt, .doc/.docx, .ppt/.pptx, .xls/.xlsx, .csv, .json, or .md files — or paste raw text straight into the editor. Audio and video (.mp3, .wav, .ogg, .m4a, .mp4, .webm) are transcribed first, so a recorded training call becomes searchable knowledge.

02

Chunk, embed, tag

Zoice splits documents into chunks, embeds them for hybrid RAG retrieval, and tags each chunk by language. Chunk health stats — total chunks, embedding coverage, average chunk size — stay visible, with an automatic "rechunk recommended" hint when the index drifts.

03

Prove it in the test bench

Before going live, run a real query in the retrieval test bench and inspect the exact top-K chunks the agent would receive — same content, same ranking. No guessing whether the right passage will surface on a live call.

04

Attach to agents

Attach knowledge bases to any agent in the editor — voice, WhatsApp, or web chat. Agents append a "Sources:" citation footer to grounded answers, and LLM-as-judge hallucination detection can be toggled per agent under Trust & Compliance.

Capabilities

A knowledge layer built for trust

Not a black-box vector store — every stage from ingestion to citation is inspectable.

Every format your knowledge lives in

Upload .pdf, .txt, .doc/.docx, .ppt/.pptx, .xls/.xlsx, .csv, .json, and .md files — or skip files entirely and paste raw text straight into a knowledge base.

Audio and video become knowledge

Drop in .mp3, .wav, .ogg, .m4a, .mp4, or .webm files. Zoice transcribes them into searchable chunks, so yesterday's recorded call can answer tomorrow's customer.

Hybrid RAG retrieval

Semantic embeddings and keyword matching work together, so exact terms like policy codes and product SKUs surface alongside loosely-phrased questions.

Per-language chunk tagging

Each chunk carries a language tag, so a Hindi question retrieves Hindi answers instead of a translation-mangled English passage — built for agents speaking 10+ Indian languages.

Chunk health, monitored

Total chunks, embedding coverage, and average chunk size at a glance — plus an automatic "rechunk recommended" hint when documents change and the index drifts.

Hallucination triage

An LLM-as-judge checks whether each answer is grounded in retrieved chunks. Ungrounded answers land in a triage dashboard with low/medium/high severity and an acknowledge/dismiss workflow.

Deep dive

Upload, test, cited answer

Walk the exact pipeline your documents travel — and see why agents stop guessing.

Drop files — or paste raw text

Audio and video are transcribed before chunking.

.pdf.docx.pptx.xlsx.csv.json.md.txt.mp3.wav.mp4.webm

Home-Loan-Policy.pdf

482 chunks · hi / en

Branch-Training-Call.mp3

Transcribed · 96 chunks

Rate-Card-Q2.xlsx

38 chunks · en

Rechunk recommended — chunk health (total chunks, embedding coverage, average chunk size) is tracked per knowledge base.

Knowledge Base — Frequently Asked Questions

Stop letting your agents guess

Upload your documents, test retrieval live, and ship agents that cite their sources on every grounded answer.

ZOICE