diff --git a/.readthedocs.yaml b/.readthedocs.yaml
new file mode 100644
index 0000000..59b3c9f
--- /dev/null
+++ b/.readthedocs.yaml
@@ -0,0 +1,16 @@
+# SPDX-License-Identifier: MIT; Copyright 2026 Andrew D Smith
+
+# Read the Docs configuration file for MkDocs projects
+# See https://docs.readthedocs.io/en/stable/config-file/v2.html for details
+
+# Required
+version: 2
+
+# Set the version of Python and other tools you might need
+build:
+  os: ubuntu-24.04
+  tools:
+    python: "3.12"
+
+mkdocs:
+  configuration: docs/readthedocs/mkdocs.yml
diff --git a/README.md b/README.md
index 24991b9..4c789da 100644
--- a/README.md
+++ b/README.md
@@ -12,7 +12,7 @@ them [here](https://github.com/smithlabcode/falco/discussions).
 Falco was conceived as an emulation of the popular
 [FastQC](https://www.bioinformatics.babraham.ac.uk/projects/fastqc) software to
 check large sequencing reads for common problems. Falco was rewritten for
-version 2.0 in order to facilitate incoroprating new functionality moving
+version 2.0 in order to facilitate incorporating new functionality moving
 forward.
 
 ## Quick start
@@ -23,7 +23,7 @@ Example:
 ```
 falco -o output input.fq
 ```
-This generates 3 files in the direcotry named `output` (creating it if needed):
+This generates 3 files in the directory named `output` (creating it if needed):
 * `fastqc_data.txt` is a text file with a summary of the QC metrics.
 * `fastqc_report.html` is the visual HTML report showing plots of the QC metrics
   summarized in the text summary.
@@ -140,20 +140,20 @@ you see anything incorrect.
   all centered values, across all tiles and across all positions.
 
 Based on the assumptions above, the grade uses an extreme value statistic. When
-introducing multithreading to falco, the order of reads analyzed changes between
+introducing multi-threading to falco, the order of reads analyzed changes between
 runs, so the 1/10 reads contributing to the tile analysis also changes between
 runs. I've noticed that this can lead to differences between runs, and in some
 cases this has changed the grade between pass/warn and warn/fail. So it is
 possible the grade can differ between runs for the same data.
 
-About input from stdin: If you are using standard input tile analysis is diabled
+About input from stdin: If you are using standard input tile analysis is disabled
 unless you specify where to find the tile info in the read name, e.g.,
 `cat file.fq | falco --stdin fq:4 outdir`, where the 4 indicates that the tile is
 after the 4th colon in the read name. The only supported positions are 4 and 6.
 
 ### Duplication results
 
-I changed how falco evaluates "duplcation". Although the format of the output is
+I changed how falco evaluates "duplication". Although the format of the output is
 the same, the numbers differ dramatically. The motivation for the change is to
 produce more useful output, and to soon build
 [preseq](https://github.com/smithlabcode/preseq) into falco.
@@ -171,7 +171,7 @@ can be turned on with `--orig-dups`. Here's how the old and new analyses differ.
 - 1,000,000 reads are hashed and counted (first 50nt of each read).
 - These are taken approximately uniformly throughout the input.
 - Although there is no randomization, if multiple threads are used the results
-  will appear as though they are randomly sampled due to fluctaions in thread
+  will appear as though they are randomly sampled due to fluctuations in thread
   speed changing which 1M reads are hashed.
 
 ## Citing falco
diff --git a/docs/readthedocs/docs/index.md b/docs/readthedocs/docs/index.md
new file mode 100644
index 0000000..03ce9fd
--- /dev/null
+++ b/docs/readthedocs/docs/index.md
@@ -0,0 +1,177 @@
+# Falco
+
+Falco v2.0 has arrived. If you want to see some recent benchmarks you can find
+them [here](https://github.com/smithlabcode/falco/discussions).
+
+Falco was conceived as an emulation of the popular
+[FastQC](https://www.bioinformatics.babraham.ac.uk/projects/fastqc) software to
+check large sequencing reads for common problems. Falco was rewritten for
+version 2.0 in order to facilitate incorporating new functionality moving
+forward.
+
+## Quick start
+
+You will be able to find binaries for Linux and macOS with the releases.
+
+Example:
+```console
+falco -o output input.fq
+```
+This generates 3 files in the directory named `output` (creating it if needed):
+
+- `fastqc_data.txt`: a text file with a summary of the QC metrics.
+- `fastqc_report.html`: the visual HTML report showing plots of the QC metrics
+  summarized in the text summary.
+- `summary.txt`: a tab-separated file describing whether the pass/warn/fail
+  result for each analysis done.
+
+If you use [anaconda](https://anaconda.org) to manage your packages, and the
+`conda` binary is in your path, you can install the most recent release of Falco
+by running
+```console
+conda install -c bioconda falco
+```
+
+## Building
+
+- Falco uses features of C++23. To compile on Linux: GCC >= 14.2.0 or LLVM-Clang
+  20.0.0 or later.  On macOS, GCC >= 15, but not GCC 16.1 which has a bug. GCC
+  16.2 is available through homebrew, so there is no reason to use 16.1
+- Falco uses the cmake build system.
+- Dependencies:
+  * [HTSLib](https://github.com/samtools/htslib): used for identifying file formats.
+  * [Zlib](https://github.com/madler/zlib): used by HTSLib for regular gzip files.
+  * [libdeflate](https://github.com/ebiggers/libdeflate): also used by HTSLib for BGZF files.
+  * [ISA-L](https://github.com/intel/isa-l) (highly recommended; not required):
+    speeds up parsing regular gzip files.
+  * Other dependencies are in the source and listed with `falco --licenses`
+- The "workflows" in the Falco GitHub repo have exactly working steps for builds
+  but are likely more than you need: `.github/workflows/linux-release.yml` and
+  `.github/workflows/macos-release.yml`
+
+From the root of the repo if you have all dependencies:
+```console
+cmake -B build -DCMAKE_CXX_COMPILER=g++ -DCMAKE_BUILD_TYPE=Release
+cmake --build build -j8  # use -j for more cores when building
+```
+Then `falco` will be in the `build` directory. Please do not build with
+`-DCMAKE_BUILD_TYPE=Build` because I wrote the source with the intention of
+allowing the compiler do most of the optimizations and without using `Release`
+falco will become slow. You can use a different compiler for the
+`-DCMAKE_CXX_COMPILER` option, but I've found a compiler often needs to be
+specified directly.
+
+### Linux detailed instructions
+
+I'm explaining this via a clean Ubuntu instance in docker:
+```console
+docker run -it ubuntu:latest bash
+```
+Inside the docker:
+```console
+export DEBIAN_FRONTEND=noninteractive &&
+apt-get update &&
+apt-get install -y --no-install-recommends \
+    zlib1g-dev \
+    libdeflate-dev \
+    libisal-dev \
+    libhts-dev \
+    ca-certificates \
+    git \
+    g++-15 \
+    cmake \
+    make \
+    samtools && # samtools is for running tests
+git clone https://github.com/smithlabcode/falco.git &&
+cd falco &&
+cmake -B build -DUSE_ISAL=on -DCMAKE_CXX_COMPILER=g++-15 -DCMAKE_BUILD_TYPE=Release &&
+cmake --build build -j8 &&  # use -j for more cores when building
+ctest --test-dir build
+```
+If you want instructions that also include building the dependencies from source,
+you can find them here: `.github/workflows/linux-release.yml`
+
+### macOS detailed instructions
+
+I don't have the same ability to test with clean OS images for macOS
+(suggestions welcome). The best I can do is use the GitHub macOS runners, which
+already have some of the dependencies installed. Here is what works:
+```console
+brew install libdeflate htslib samtools &&  # samtools for testing
+git clone https://github.com/smithlabcode/falco.git &&
+cd falco &&
+cmake -B build -DCMAKE_CXX_COMPILER=g++-15 -DCMAKE_BUILD_TYPE=Release &&
+cmake --build build -j8 &&
+ctest --test-dir build
+```
+ZLib is already installed on macOS, HTSLib installs libdeflate as a dependency
+and samtools installs both as dependency. To see what's already installed on
+GitHub's macOS look
+[here](https://github.com/actions/runner-images/blob/main/images/macos/macos-26-Readme.md).
+Note: ISA-L was designed for Intel hardware. It works on Apple silicon, but only
+gives 1-2% speedup in my tests.
+
+## Changes in Falco v2.0
+
+### Tiles results
+
+I found that the method for tile analysis is a bit unstable, and the tile grade
+can be slightly unstable. The only way to notice this is to process reads from
+the same input file in different orders, which happens as a side effect of
+analyzing reads concurrently with threads. Here is my understanding of how
+FastQC works, and how I implemented tile analysis in falco v2. Please comment if
+you see anything incorrect.
+
+- Tile analysis is done for 1/10 of the reads (though FastQC includes all among
+  the first 10k reads).
+- If more than 2500 tiles are identified, an error is assumed.
+- Accumulating results: For each counted read, for each position in the read,
+  the quality score contributes to that tile's mean for the given position.
+- Summarizing tile results: For each read position, the mean over tiles' quality
+  scores is taken. Then for each tile, for each read position, the value is
+  centered by subtracting the mean (the 'centered' tile values).
+- The summary stat for deciding the grade is based on the minimum value among
+  all centered values, across all tiles and across all positions.
+
+Based on the assumptions above, the grade uses an extreme value statistic. When
+introducing multi-threading to falco, the order of reads analyzed changes between
+runs, so the 1/10 reads contributing to the tile analysis also changes between
+runs. I've noticed that this can lead to differences between runs, and in some
+cases this has changed the grade between pass/warn and warn/fail. So it is
+possible the grade can differ between runs for the same data.
+
+About input from stdin: If you are using standard input tile analysis is disabled
+unless you specify where to find the tile info in the read name, e.g.,
+`cat file.fq | falco --stdin fq:4 outdir`, where the 4 indicates that the tile is
+after the 4th colon in the read name. The only supported positions are 4 and 6.
+
+### Duplication results
+
+I changed how falco evaluates "duplication". Although the format of the output is
+the same, the numbers differ dramatically. The motivation for the change is to
+produce more useful output, and to soon build
+[preseq](https://github.com/smithlabcode/preseq) into falco.
+
+The original duplication analysis method is still implemented in falco v2.0, and
+can be turned on with `--orig-dups`. Here's how the old and new analyses differ.
+
+#### Original method
+- 100,000 unique reads are hashed and counted (first 50nt of each read).
+- These are taken from the first 100,000 reads in the input.
+- After the first 100,000 unique has been reached, reads that match one
+  previously hashed are counted.
+
+#### New method
+- 1,000,000 reads are hashed and counted (first 50nt of each read).
+- These are taken approximately uniformly throughout the input.
+- Although there is no randomization, if multiple threads are used the results
+  will appear as though they are randomly sampled due to fluctuations in thread
+  speed changing which 1M reads are hashed.
+
+## Citing falco
+
+If falco was helpful for your research, you can cite us as follows:
+
+de Sena Brandine G and Smith AD. Falco: high-speed FastQC emulation for quality
+control of sequencing data. F1000Research 2021, 8:1874
+(https://doi.org/10.12688/f1000research.21142.2)
diff --git a/docs/readthedocs/mkdocs.yml b/docs/readthedocs/mkdocs.yml
new file mode 100644
index 0000000..cd26b3d
--- /dev/null
+++ b/docs/readthedocs/mkdocs.yml
@@ -0,0 +1,6 @@
+site_name: falco
+strict: true
+
+theme: readthedocs
+nav:
+   - Home: 'index.md'