One inbox. Every channel.
Widget chat, email, and AI replies converge into a single team queue with assignments, SLAs, and CSAT.
Hybrid AI knowledge core. Public docs + private uploads. One engine.
Your knowledge, hybrid. Your chatbot, instant.
Mix what's public on your website with what's private in your team's docs. AskCore blends both into a single AI core and serves the right answers in every conversation — public knowledge for visitors, private knowledge for support flows, all from one platform.
Most chatbot vendors force you to choose
Public-only (crawl my site) or private-only (upload my docs). Neither matches how real support actually works.
Public-only bots stay shallow
Crawling marketing pages alone misses the runbooks, internal wikis, and product detail your team actually uses.
Private-only bots stay siloed
Upload-only chatbots can't answer the basic questions your homepage already covers — visitors hit a wall on day one.
Two tools means two bills, two integrations, and conflicting answers
Operators end up reading from two systems. Customers get inconsistent replies across channels.
AskCore is the hybrid: one AI core, both source types, scoped per chatbot, served everywhere.
One inbox. One AI core. Every channel.
Widget chat, email, and AI replies converge into a single team queue. Powered by hybrid knowledge that cites every answer.
Hybrid knowledge. Real citations.
Mix public sources (sitemap, URL crawl, help center) with private uploads (ZIPs, paste-in text). Every answer cites its source.
Built for teams, not just demos.
Per-operator chatbot access, per-source scoping, audited shadow sessions. Predictable, reviewable, hard to misuse.
What a hybrid AI core looks like in production
Public knowledge for visitors. Private knowledge for support-only flows. Operators see what they own. Audits cover everything.
Widget + Email + Help center
one customer conversation, three channels, same answers
Public + Private RAG
scope each knowledge source to all chatbots or selected ones
Audit + Shadow + Reports
reproduce any user's view, log every action, measure CSAT separately for AI and humans
Sharded ZIP indexing, automatic recovery from rate-limited embedding calls, and self-healing for interrupted jobs — all built in. Your re-index will not get stuck overnight.
See the hybrid flow before you put it live
Start with an internal test page. Watch a public-source answer cite the homepage. Watch a private-source answer cite an internal ZIP. Tune both before flipping the widget on.
Install the widget when you are ready to go live
<script src="https://askcore.org/widget.js" data-tenant="your-tenant" data-chatbot="support" data-api-key="wb_public_key"></script>From safe testing to go-live in three steps
Start internally with both source types. Set up the workspace. Go live when the team is ready.
1
Test with both source types
Upload a private ZIP, point AskCore at your sitemap, and ask the AI questions on the internal test page. See exactly which source each answer comes from.
2
Set up your workspace
Invite your team, scope each chatbot to its public + private sources, restrict operators to their queue, and configure SLAs and automation.
3
Go live on your website
Install the widget last. Public sources serve visitors. Private sources stay invisible to anyone outside the chatbots you've scoped them to.
Stop reading. Try it on your own docs.
Sign up, paste a sitemap URL, drag in a ZIP, and ask it a question. If the answer is good, you are 30 minutes from production. If it is not, you will see exactly which doc to improve.