Counting annual-scale fluctuations, such as geochemical cyclicity or visible growth bands, within a climate archive can yield extremely high-precision chronological models. However, this process is often time-consuming and subjective, and although various software packages can automate this process, many researchers still prefer to count manually given its technical simplicity and transparency. Here we present a new tool that combines the time saved by automation with the flexibility afforded by expert judgement. CYCLIM uses a matched filtering approach to detect cyclicity and then allows the user to inspect and refine the automated output whilst also quantifying age uncertainty. The presented framework speeds up cycle counting by automating the first-pass of the count while also retaining the benefits of a manual count by allowing for post-analysis tuning. Across three examples using published palaeoclimate reconstructions, the automatic output found 96.2 % of the cycles, with a false positive and false negative rate of 3.3 % and 3.8 %, respectively. This means that only 7 cycles per 100 need to be corrected manually, making cycle counting with CYCLIM ~14.1 times faster.