Kelu vs Chatbase
Last reviewed:
Both let you build an AI chatbot that answers from your content. Chatbase is a general AI agent platform for customer experience across chat, email, voice and messaging apps. Kelu is purpose-built for docs and support, with hybrid retrieval, citations and abstention: the pieces that stop the assistant guessing when it doesn't know.
Choose Chatbase if you need one agent across WhatsApp, Instagram, Messenger, phone and email, or native iOS and Android SDKs. Choose Kelu if you want answers for docs and support that cite their sources and say so when your docs don't cover a question, plus Slack and Teams bots, an MCP server and Support Ticket AI in your helpdesk, with a free plan to start.
Feature comparison
Chatbase details come from Chatbase's own website, pricing page and docs, listed under the table.
Where Kelu focuses differently
Built for docs, not for everything
Chatbase spans support, sales, voice, and WhatsApp: it's a general platform. Kelu is deliberately narrower: it ingests your docs, wikis, tickets, and code, and answers with citations. That focus is why the retrieval stack (hybrid + reranker + abstention) exists and why the docs-platform install guides (Mintlify, Docusaurus, GitBook, ReadMe, VitePress) are first-class.
The assistant refuses to guess
An abstention gate withholds your content from the model when retrieval finds the question isn't covered, so the answer becomes 'I don't have that in the docs' rather than a plausible-sounding invention. Combined with grounded citations, this is what makes Kelu safe to put on a customer-facing widget.
More surfaces bundled
Slack and Microsoft Teams bots, an MCP server for AI coding tools, a support form deflector and Support Ticket AI for your helpdesk, all included on Pro. No add-on fees.
Open API and MCP
The hosted widget, an open REST API with streaming, and an MCP server per knowledge base, on every plan. Build custom experiences without waiting for a native integration. React and JavaScript SDKs are on npm.
Got questions?
Frequently asked questions
Is Kelu a good replacement for Chatbase?
If your use case is docs and support with a widget, Slack or Teams bots, helpdesk answers, or an MCP server for AI coding tools, Kelu is purpose-built for it, with citations and an abstention gate that Chatbase doesn't document. If you need WhatsApp, Instagram, phone or native mobile SDKs, Chatbase covers channels Kelu doesn't.
How does retrieval quality compare?
Kelu combines vector (pgvector) with PostgreSQL full-text search, fuses them with reciprocal rank fusion, rewrites the query with an LLM before search, and reranks with a cross-encoder. Vector-only retrieval misses exact-match queries (error codes, SKUs). Chatbase does not describe its retrieval publicly.
Does every answer include citations?
Every answer built from your content does: citations link to the sources it used. A reply to small talk carries none, and when your sources don't cover a question, the assistant abstains instead of generating a plausible-sounding answer.
Can I move from Chatbase to Kelu?
Yes. Reconnect the same sources (website URLs, files, Notion), swap the widget embed for Kelu's script tag, and you're running. A first cited answer takes minutes once a docs source has synced.
How does pricing compare?
Kelu has a free plan (100 questions/mo), Pro at $99/mo or $79/mo billed yearly (1,000 questions, 5 knowledge bases), and Enterprise custom. Chatbase's plans are Hobby $32, Basic $120 and Business $400 a month billed yearly, with a 7-day trial and no free plan on its pricing page.
See for yourself
Connect your docs and get your first AI-powered answer in under 5 minutes. No credit card required.
Free to start · No credit card required · Talk to Sales →