Frequently Asked Questions
Everything you need to know about Kelu.
Also useful: all features · alternatives · compare · trust center · free tools.
General
What is Kelu?
Kelu is an AI Knowledge Assistant. It connects your docs, wikis, policies, product content, code repos, community forums, and helpdesk knowledge — then answers questions with citations on every surface: a hosted widget, Slack and Microsoft Teams bots, a browser extension, an MCP server for AI IDEs, Support Ticket AI inside your helpdesk, and a public REST API.
How is Kelu different from just using ChatGPT?
ChatGPT answers from its training data, which doesn't include your product documentation, internal policies, or private knowledge. Kelu retrieves from your indexed content and cites the exact page each answer came from — so teams and customers can verify the response, and the assistant abstains when it doesn't have grounded evidence rather than making something up.
Do I need machine-learning expertise to use Kelu?
No. You connect a source (a URL, a GitHub repo, a Notion workspace, a helpdesk) and Kelu handles chunking, embedding, indexing, retrieval, reranking, and generation. Nothing to fine-tune, no infrastructure to run.
How does Kelu compare to Kapa / Inkeep / Chatbase / Intercom Fin?
We have detailed side-by-side comparisons at /alternatives (26 competitor pages) and /compare (category-level round-ups). The short version: Kelu ships more surfaces on one plan (widget + Slack + Teams + MCP + browser extension + Support Ticket AI + form deflector), has a real self-serve free tier, and combines hybrid retrieval (vector + full-text + reranker) with an enforced abstention gate.
Is Kelu a good alternative to Kapa.ai?
Kelu and Kapa.ai both answer questions from technical documentation with citations. Kelu also answers helpdesk tickets in Zendesk, Intercom, Freshdesk, Front and Linear, answers support-form questions before a ticket is filed, and includes Slack and Teams bots and an MCP server. Kelu has a free plan, and Pro costs $99 a month. See /alternatives/kapa and /compare/kapa-vs-inkeep-vs-kelu.
RAG, Citations & Grounding
What is RAG and why does it matter?
RAG (Retrieval-Augmented Generation) retrieves relevant documents first, then generates an answer grounded in those documents. This sharply reduces hallucinations about your specific product and makes answers verifiable via citation links. See /glossary/rag for a longer explanation.
Does Kelu cite its sources?
Yes. Every answer from your docs carries numbered links to the pages it came from. If your content doesn't cover the question, Kelu says it doesn't know and shows no sources.
How are citations generated?
During retrieval, each source chunk is tracked with its URL and title. The generation prompt requires the model to ground in those sources. Every response includes a structured citations array with the URL, title, and excerpt for every source chunk used — the widget, bots, MCP, and API all return the same citations.
What happens when Kelu can't find a good answer?
When retrieval finds that your sources do not cover the question, the assistant abstains — it gets no excerpts and no citations, and says it cannot answer instead of guessing. This is enforced by an abstention gate, not just prompted. Unanswered questions become part of the gap report and can be turned into Doc Proposals (AI-drafted answers your team reviews and adds to Q&A).
How does hybrid search work?
Kelu combines semantic vector search (pgvector cosine) with Postgres full-text ranking. Results are merged with reciprocal rank fusion. Before search, an LLM rewrites the query to improve recall. After retrieval, a cross-encoder reranker sorts the top candidates. See /features/hybrid-search for the full pipeline.
How do you measure answer quality?
Readers rate answers thumbs up or down, analytics show how many answers cite a source, and the gap report lists the questions your sources don't cover. Together you see quality as a moving number, not a demo. See /features/evals-quality.
Surfaces & Integrations
Where can Kelu answer questions?
Hosted web widget, Slack and Microsoft Teams bots (both native), a browser extension (Chrome / Edge / Opera MV3 side panel), an MCP server per knowledge base for Cursor / Claude Desktop / ChatGPT with MCP, a standalone form deflector for support-form ticket deflection, Support Ticket AI inside Zendesk, Intercom, Freshdesk, Front and Linear, a public REST + WebSocket API, and React and JavaScript SDKs on npm.
Can Kelu index GitHub repositories?
Yes. Connect a GitHub repository and specify which paths to index (README.md, docs/, ADRs, changelogs). A push webhook triggers incremental re-index on every commit to the configured branch. GitHub issues, pull requests and discussions are separate source types you can index alongside the repository.
Which helpdesks does Support Ticket AI integrate with?
Zendesk, Intercom, Freshdesk, Front, and Linear for tickets, plus Freshchat for live chat. Zendesk Messaging is coming soon. Jira Service Management and Salesforce Service Cloud are paused for new connections. On tickets, each integration can leave AI drafts as private notes inside the helpdesk (Draft only mode, not on Linear, where a note is a public comment), or auto-reply above your confidence threshold. Zendesk agents also get a sidebar that drafts on request; an Intercom sidebar comes later. See /features/support-ticket-ai.
How does ticket deflection work?
Two paths. Before submit: the /features/support-form-agent form deflector answers the question when the customer submits a support form — confidence-scored, and only deflects when it has grounded evidence. In the helpdesk: Support Ticket AI reads incoming tickets and, above your confidence threshold, either drafts a reply for the agent (Draft only) or auto-responds; below it, the ticket gets the escalation tag. Simulations let you calibrate the threshold on historical tickets before enabling anything customer-facing.
Can I test Support Ticket AI before enabling it in production?
Yes — that's what Simulations do. Pick up to 100 of your most recent resolved tickets, the assistant answers each as if it were live, and you review the results in the dashboard with human sign-off per answer. Each draft sits beside the agent's real reply with its confidence score, so you can see where to set the threshold before launch. See /features/simulations.
What is the MCP server and how do I use it?
Each knowledge base can have its own Model Context Protocol (MCP) servers. Open the knowledge base, go to Integrations → MCP Server, create a server (its address is https://app.kelu.dev/mcp/<name>) and copy the install command for your tool. A non-internal knowledge base is also reachable at POST /mcp with any of its client keys that has Allow MCP access, sent as the X-Client-Key header. Cursor, Claude Desktop, ChatGPT (with MCP), Windsurf, and any MCP-compatible agent can invoke the search tool (named after the knowledge base; search_docs also works), ask_question, list_sources, and feedback as structured tools. See /features/mcp-server.
Does Kelu work with Cursor and Claude?
Yes. Each knowledge base can have its own MCP server, so Cursor, Claude Desktop and other MCP tools can search your docs while developers code. See /features/mcp-server.
Can Kelu answer Zendesk tickets automatically?
Yes. It answers the tickets it has evidence for and tags the rest for your team. In Draft only mode, confident answers become private notes instead of replies. Simulations replay past tickets first, so you can see what it would have said. See /features/support-ticket-ai.
How quickly can I put Kelu on my docs site?
Enter your docs URL so Kelu can crawl it, test some questions in the Playground, then paste one script tag on your site. The steps are at /docs/quickstart.
What's the browser extension?
A Chrome / Edge / Opera MV3 side-panel extension that opens your knowledge base next to whatever tab you're on. It is not in the Chrome Web Store yet, so there is no self-serve install today. Ask questions about the page, look up an API without leaving your editor, or pull an internal policy answer without breaking flow. Signs in with your Kelu account and respects the same permissions as the dashboard. See /features/browser-extension.
How do the Slack / Teams bots work?
Installed per workspace from the dashboard, on Pro and above. Slack: /kelu slash command + @mention + forum auto-reply. Teams: mention-based Q&A. Discord bot: switched off for now. Both carry citations, thumbs feedback (up/down), and identical answer quality — same retrieval pipeline as the widget and API.
What are Collections and per-collection permissions?
Collections (source groups) let you scope retrieval by product line, version, or region. Each widget, support form deflector, and Slack or Teams bot can be set to answer only from the groups you pick, and a page's own setting can only narrow that, never widen it. Source groups are on Enterprise.
Does Kelu support versioned documentation?
Yes. Give each source a version label (v1, v2, latest), put each version in its own source group, and scope each docs version's widget to its group. Users then get answers from the docs version they are reading. Source groups are on Enterprise.
Doc Proposals
What are Doc Proposals?
Gap analytics logs every unanswered question and ranks the missing knowledge by volume. Doc Proposals (Knowledge Maintenance in the dashboard) drafts an answer from your own content — your team fills any gaps and adds it to Q&A, or dismisses it. AI proposes, humans publish. See /features/doc-proposals.
Security, Privacy & Data
Is my content used to train AI models?
No. Indexed content, embeddings, and conversations stay isolated to your workspace and are used only to answer your own questions. Nothing is shared across customers or used to train any model — ours or a third party's. This is contractual on paid plans and technical on every plan.
How is PII handled?
When you turn it on for a knowledge base, PII (emails, phone numbers, credit-card numbers, IBANs, IP addresses, names after a title such as Dr, and your own patterns) is masked in indexed content and in widget, API and Ask questions — before the LLM call and before storage. Questions from bots, the form deflector, helpdesk automation, MCP and the search API are not masked. See /features/pii-and-zero-retention.
What is zero-retention mode?
A per-workspace toggle. Questions and answers stream to the caller in real time and are then dropped — no conversation row is created, no message is stored. Retrieval still runs against your indexed content and the answer is still cited. Only the transcript is not persisted.
Can I use my own OpenAI / Anthropic / Gemini key (BYOK)?
Yes. BYOK routes LLM traffic through your provider account with your key, under your existing data-processing agreement. You keep the model choice and the invoice; Kelu is the platform layer. See /features/byok.
Is Kelu SOC 2 certified?
Not yet. Our SOC 2 Type II audit is in progress, so there is no report to share today. A DPA is at /dpa, and /trust shows the current control posture.
Where is my data stored?
In the United States. Data is encrypted at rest (AES-256) and in transit (TLS 1.3).
Do you support SSO?
Yes on Enterprise: OIDC and SAML 2.0. Works with Okta, Entra (Azure AD), Google Workspace, OneLogin, and any SAML-compliant IdP.
Is the widget protected against abuse?
Yes. The hosted widget at widget.kelu.dev is protected by reCAPTCHA to prevent automated bots from consuming your question quota. Runs invisibly; no user friction. Server-side calls with a client ID + secret pair skip it.
How are client keys secured?
Public client keys (kl_pk_...) are safe to embed in client-side code — they're scoped to a single knowledge base and only allow public API calls (chat, search, form deflection, feedback) plus MCP unless you untick Allow MCP access when you create the key. Backend integrations use a client ID + secret pair (kl_ci_... / kl_cs_...) which must not be exposed client-side. Keys are revocable instantly from the dashboard.
Pricing & Limits
Is there a free plan?
Yes. The Free plan includes 1 knowledge base, up to 500 indexed pages, 100 AI questions per month, and 1 team member — no credit card required.
How much is Pro?
$99/month billed monthly, or $79/month billed yearly. Includes 5 knowledge bases, up to 10,000 indexed pages, 1,000 AI questions per month, up to 10 team members, Slack / Teams bots, and Support Ticket AI. MCP and analytics are on every plan.
What counts as an AI question?
Each end-user message the assistant answers counts as one question — whether it comes through the widget, a bot (Slack / Teams), the browser extension, Support Ticket AI, MCP, or the public API. Follow-up messages in the same conversation each count as a question. Search API queries (retrieval without generation) do not count.
Can I try Pro before paying?
Yes. Pro comes with a 14-day free trial. Stripe Checkout collects your card to start it, and the trial is for a workspace's first subscription only.
What happens if I exceed my monthly question limit?
Owners and admins are emailed at 50, 80, 90, 95 and 100% of your quota. Once you hit 100%, new questions are rejected with a specific 'monthly quota exceeded' response — a hard limit, not a suggestion, so you don't get a surprise LLM bill. Upgrading takes effect immediately; otherwise the quota resets at the start of the next calendar month.
Do actions or MCP calls count against my quota?
Only questions Kelu answers count against the AI-question quota, including form deflections and helpdesk tickets. MCP list_sources and the search tool (retrieval-only) don't count; ask_question calls do.
What's on Enterprise?
Custom quotas, SSO (OIDC / SAML), a countersigned DPA, source groups, hide-branding on the widget, dedicated support, custom SLA. Contact us for pricing.