Platforms

How each platform's knowledge base feature works, where it goes wrong on calls, and how to load your knowledge base into it. We are not affiliated with any of them.

• The Retell AI knowledge base: how it works, and how to load your knowledge base into it

  How Retell AI's knowledge base works: automatic retrieval every turn, 2,000-character chunks, 3 chunks by default, the knowledge base instruction, cost, and how to load a Markdown knowledge base.

• The Vapi knowledge base: how it works, and how to load your knowledge base into it

  How Vapi's knowledge base works: files behind a query tool the assistant must decide to call, the 300 KB advice, how to update files, and how to load a Markdown knowledge base so the assistant doesn't mix up prices.

• The Botpress knowledge base: how it works, and how to load your knowledge base into it

  How Botpress knowledge bases work: the Search Knowledge card, accepted files including Markdown, chunks count, updating documents, and how to load a knowledge base file so the bot doesn't mix up prices.

• The Voiceflow knowledge base: how it works, and how to load your knowledge base into it

  How Voiceflow's knowledge base works: 3 chunks per query by default, chunk size, LLM chunking strategies, refresh and publishing, and how to load a knowledge base file so the agent doesn't mix up prices.

• The Bland AI knowledge base: how it works, and how to load your knowledge base into it

  How Bland AI's Knowledge sources work: the agent decides whether to search from each source's name and description, what you can upload, limits, and how to load a knowledge base file so the agent doesn't mix up prices.

• The Bolna knowledge base: how it works, and how to load your knowledge base into it

  How Bolna's knowledge base works for AI calling agents: PDF or URL only, RAG on every turn, chunk and top-k settings, and how to load a knowledge base file so the agent doesn't mix up prices.

• The Synthflow knowledge base: how it works, and how to load your knowledge base into it

  How Synthflow's knowledge base works: sources, the 'when to search' condition, RAG lookups, updating sources, and how to load a knowledge base file so the agent doesn't mix up prices.

• The ElevenLabs Agents knowledge base: how it works, and how to load your knowledge base into it

  How ElevenLabs Agents uses a knowledge base: prompt mode versus RAG, the 300,000-character limit, RAG settings, and how to load a Markdown knowledge base so your agent doesn't mix up prices.