CCBot — Your intelligent content assistant
CCBot understands your PDFs, web pages and Drupal content and responds with verifiable answers. Each answer includes a citation back to the source—down to the paragraph or anchor—so users can check context and explore further.
What it can do
- • Cited answers that link to precise paragraphs/anchors in PDFs and web pages.
- • Memory across turns (session by default; optional persistent memory per tenant).
- • Multilingual queries and answers in EN/FR/DE/ES/NL; cross‑lingual retrieval.
- • Tables & long docs — extracts facts from tables and annexes across thousands of pages.
- • Embeddable via Drupal module today; JS/iFrame embed on the roadmap.
- • Admin controls for sources, context size, throttling, and allow/deny lists.
Safety & governance
- • Jailbreak resistance with layered prompt hardening and output filtering.
- • PII protection, policy blocklists and rate limiting for abuse prevention.
- • Monitoring with configurable thresholds and incident workflows.
- • Human oversight: conversation logs; export for DPIAs and audits.
Performance & scale
Optimised for thousands of PDFs and large sites. Typical answers return in ~1–3 seconds depending on context size and load. Caching and streaming keep interactions responsive.
Data protection
- • EU data residency; encryption in transit and at rest.
- • Admin‑set retention with export/delete on request.
- • DPA available; sub‑processors listed in the Trust Center.
FAQ
Which sources are supported? PDFs and HTML (static/dynamic) at launch; Drupal nodes/REST; more sources coming.
Can we restrict topics? Yes. Use allow/deny lists and per‑source scoping; configure maximum context and response policies.
How do languages work? We support cross‑lingual retrieval and answers in EN/FR/DE/ES/NL; we can add more EU languages on request.