diff --git a/src/backup.rs b/src/backup.rs
index c05b423..ba7a76b 100644
--- a/src/backup.rs
+++ b/src/backup.rs
@@ -1,11 +1,11 @@
 //! `feature = "backup"` Online SQLite backup API.
 //!
-//! To create a `Backup`, you must have two distinct `Connection`s - one
+//! To create a [`Backup`], you must have two distinct [`Connection`]s - one
 //! for the source (which can be used while the backup is running) and one for
-//! the destination (which cannot).  A `Backup` handle exposes three methods:
-//! `step` will attempt to back up a specified number of pages, `progress` gets
-//! the current progress of the backup as of the last call to `step`, and
-//! `run_to_completion` will attempt to back up the entire source database,
+//! the destination (which cannot).  A [`Backup`] handle exposes three methods:
+//! [`step`](Backup::step) will attempt to back up a specified number of pages, [`progress`](Backup::progress) gets
+//! the current progress of the backup as of the last call to [`step`](Backup::step), and
+//! [`run_to_completion`](Backup::run_to_completion) will attempt to back up the entire source database,
 //! allowing you to specify how many pages are backed up at a time and how long
 //! the thread should sleep between chunks of pages.
 //!
@@ -130,7 +130,7 @@ impl Connection {
     }
 }
 
-/// `feature = "backup"` Possible successful results of calling `Backup::step`.
+/// `feature = "backup"` Possible successful results of calling [`Backup::step`].
 #[derive(Copy, Clone, Debug, PartialEq, Eq)]
 #[non_exhaustive]
 pub enum StepResult {
@@ -152,8 +152,8 @@ pub enum StepResult {
 
 /// `feature = "backup"` Struct specifying the progress of a backup. The
 /// percentage completion can be calculated as `(pagecount - remaining) /
-/// pagecount`. The progress of a backup is as of the last call to `step` - if
-/// the source database is modified after a call to `step`, the progress value
+/// pagecount`. The progress of a backup is as of the last call to [`step`](Backup::step) - if
+/// the source database is modified after a call to [`step`](Backup::step), the progress value
 /// will become outdated and potentially incorrect.
 #[derive(Copy, Clone, Debug)]
 pub struct Progress {
@@ -225,7 +225,7 @@ impl Backup<'_, '_> {
         })
     }
 
-    /// Gets the progress of the backup as of the last call to `step`.
+    /// Gets the progress of the backup as of the last call to [`step`](Backup::step).
     #[inline]
     pub fn progress(&self) -> Progress {
         unsafe {
@@ -240,7 +240,7 @@ impl Backup<'_, '_> {
     /// negative, will attempt to back up all remaining pages. This will hold a
     /// lock on the source database for the duration, so it is probably not
     /// what you want for databases that are currently active (see
-    /// `run_to_completion` for a better alternative).
+    /// [`run_to_completion`](Backup::run_to_completion) for a better alternative).
     ///
     /// # Failure
     ///
@@ -262,7 +262,7 @@ impl Backup<'_, '_> {
         }
     }
 
-    /// Attempts to run the entire backup. Will call `step(pages_per_step)` as
+    /// Attempts to run the entire backup. Will call [`step(pages_per_step)`](Backup::step) as
     /// many times as necessary, sleeping for `pause_between_pages` between
     /// each call to give the source database time to process any pending
     /// queries. This is a direct implementation of "Example 2: Online Backup
@@ -276,7 +276,7 @@ impl Backup<'_, '_> {
     ///
     /// # Failure
     ///
-    /// Will return `Err` if any of the calls to `step` return `Err`.
+    /// Will return `Err` if any of the calls to [`step`](Backup::step) return `Err`.
     pub fn run_to_completion(
         &self,
         pages_per_step: c_int,
diff --git a/src/blob/mod.rs b/src/blob/mod.rs
index 2d38185..34febde 100644
--- a/src/blob/mod.rs
+++ b/src/blob/mod.rs
@@ -47,7 +47,7 @@
 //!    functions take a `&mut [MaybeUninit<u8>]` as the destination buffer,
 //!    where the "normal" functions take a `&mut [u8]`.
 //!
-//!    Using `MaybeUninit` here can be more efficient in some cases, but is
+//!    Using `MaybeUninit` here can be more efficient in some cases, but is
 //!    often inconvenient, so both are provided.
 //!
 //! 2. Exact/inexact refers to to whether or not the entire buffer must be
diff --git a/src/busy.rs b/src/busy.rs
index 39606c6..fd46c74 100644
--- a/src/busy.rs
+++ b/src/busy.rs
@@ -19,7 +19,7 @@ impl Connection {
     ///
     /// There can only be a single busy handler for a particular database
     /// connection at any given moment. If another busy handler was defined
-    /// (using `busy_handler`) prior to calling this routine, that other
+    /// (using [`busy_handler`](Connection::busy_handler)) prior to calling this routine, that other
     /// busy handler is cleared.
     pub fn busy_timeout(&self, timeout: Duration) -> Result<()> {
         let ms: i32 = timeout
@@ -45,7 +45,7 @@ impl Connection {
     ///
     /// There can only be a single busy handler defined for each database
     /// connection. Setting a new busy handler clears any previously set
-    /// handler. Note that calling `busy_timeout()` or evaluating `PRAGMA
+    /// handler. Note that calling [`busy_timeout()`](Connection::busy_timeout) or evaluating `PRAGMA
     /// busy_timeout=N` will change the busy handler and thus
     /// clear any previously set busy handler.
     pub fn busy_handler(&self, callback: Option<fn(i32) -> bool>) -> Result<()> {
diff --git a/src/cache.rs b/src/cache.rs
index 478a5e1..ea00019 100644
--- a/src/cache.rs
+++ b/src/cache.rs
@@ -11,7 +11,7 @@ impl Connection {
     /// Prepare a SQL statement for execution, returning a previously prepared
     /// (but not currently in-use) statement if one is available. The
     /// returned statement will be cached for reuse by future calls to
-    /// `prepare_cached` once it is dropped.
+    /// [`prepare_cached`](Connection::prepare_cached) once it is dropped.
     ///
     /// ```rust,no_run
     /// # use rusqlite::{Connection, Result};
@@ -63,7 +63,7 @@ pub struct StatementCache(RefCell<LruCache<Arc<str>, RawStatement>>);
 /// Cacheable statement.
 ///
 /// Statement will return automatically to the cache by default.
-/// If you want the statement to be discarded, call `discard()` on it.
+/// If you want the statement to be discarded, call [`discard()`](CachedStatement::discard) on it.
 pub struct CachedStatement<'conn> {
     stmt: Option<Statement<'conn>>,
     cache: &'conn StatementCache,
@@ -105,7 +105,7 @@ impl CachedStatement<'_> {
     }
 
     /// Discard the statement, preventing it from being returned to its
-    /// `Connection`'s collection of cached statements.
+    /// [`Connection`]'s collection of cached statements.
     #[inline]
     pub fn discard(mut self) {
         self.stmt = None;
diff --git a/src/error.rs b/src/error.rs
index 22fd970..49778a8 100644
--- a/src/error.rs
+++ b/src/error.rs
@@ -43,11 +43,11 @@ pub enum Error {
     /// Error converting a file path to a string.
     InvalidPath(PathBuf),
 
-    /// Error returned when an `execute` call returns rows.
+    /// Error returned when an [`execute`](crate::Connection::execute) call returns rows.
     ExecuteReturnedResults,
 
     /// Error when a query that was expected to return at least one row (e.g.,
-    /// for `query_row`) did not return any.
+    /// for [`query_row`](crate::Connection::query_row)) did not return any.
     QueryReturnedNoRows,
 
     /// Error when the value of a particular column is requested, but the index
@@ -67,29 +67,29 @@ pub enum Error {
     /// any or insert many.
     StatementChangedRows(usize),
 
-    /// Error returned by `functions::Context::get` when the function argument
+    /// Error returned by [`functions::Context::get`](crate::functions::Context::get) when the function argument
     /// cannot be converted to the requested type.
     #[cfg(feature = "functions")]
     InvalidFunctionParameterType(usize, Type),
-    /// Error returned by `vtab::Values::get` when the filter argument cannot
+    /// Error returned by [`vtab::Values::get`](crate::vtab::Values::get) when the filter argument cannot
     /// be converted to the requested type.
     #[cfg(feature = "vtab")]
     InvalidFilterParameterType(usize, Type),
 
     /// An error case available for implementors of custom user functions (e.g.,
-    /// `create_scalar_function`).
+    /// [`create_scalar_function`](crate::Connection::create_scalar_function)).
     #[cfg(feature = "functions")]
     #[allow(dead_code)]
     UserFunctionError(Box<dyn error::Error + Send + Sync + 'static>),
 
-    /// Error available for the implementors of the `ToSql` trait.
+    /// Error available for the implementors of the [`ToSql`](crate::types::ToSql) trait.
     ToSqlConversionFailure(Box<dyn error::Error + Send + Sync + 'static>),
 
     /// Error when the SQL is not a `SELECT`, is not read-only.
     InvalidQuery,
 
     /// An error case available for implementors of custom modules (e.g.,
-    /// `create_module`).
+    /// [`create_module`](crate::Connection::create_module)).
     #[cfg(feature = "vtab")]
     #[allow(dead_code)]
     ModuleError(String),
@@ -98,8 +98,8 @@ pub enum Error {
     #[cfg(feature = "functions")]
     UnwindingPanic,
 
-    /// An error returned when `Context::get_aux` attempts to retrieve data
-    /// of a different type than what had been stored using `Context::set_aux`.
+    /// An error returned when [`Context::get_aux`](crate::functions::Context::get_aux) attempts to retrieve data
+    /// of a different type than what had been stored using [`Context::set_aux`](crate::functions::Context::set_aux).
     #[cfg(feature = "functions")]
     GetAuxWrongType,
 
diff --git a/src/functions.rs b/src/functions.rs
index bc1de96..8482f7b 100644
--- a/src/functions.rs
+++ b/src/functions.rs
@@ -126,7 +126,7 @@ impl Context<'_> {
     ///
     /// # Failure
     ///
-    /// Will panic if `idx` is greater than or equal to `self.len()`.
+    /// Will panic if `idx` is greater than or equal to [`self.len()`](Context::len).
     ///
     /// Will return Err if the underlying SQLite type cannot be converted to a
     /// `T`.
@@ -156,16 +156,16 @@ impl Context<'_> {
     ///
     /// # Failure
     ///
-    /// Will panic if `idx` is greater than or equal to `self.len()`.
+    /// Will panic if `idx` is greater than or equal to [`self.len()`](Context::len).
     #[inline]
     pub fn get_raw(&self, idx: usize) -> ValueRef<'_> {
         let arg = self.args[idx];
         unsafe { ValueRef::from_value(arg) }
     }
 
-    /// Fetch or insert the the auxilliary data associated with a particular
+    /// Fetch or insert the auxilliary data associated with a particular
     /// parameter. This is intended to be an easier-to-use way of fetching it
-    /// compared to calling `get_aux` and `set_aux` separately.
+    /// compared to calling [`get_aux`](Context::get_aux) and [`set_aux`](Context::set_aux) separately.
     ///
     /// See `https://www.sqlite.org/c3ref/get_auxdata.html` for a discussion of
     /// this feature, or the unit tests of this module for an example.
@@ -206,7 +206,7 @@ impl Context<'_> {
     }
 
     /// Gets the auxilliary data that was associated with a given parameter via
-    /// `set_aux`. Returns `Ok(None)` if no data has been associated, and
+    /// [`set_aux`](Context::set_aux). Returns `Ok(None)` if no data has been associated, and
     /// Ok(Some(v)) if it has. Returns an error if the requested type does not
     /// match.
     pub fn get_aux<T: Send + Sync + 'static>(&self, arg: c_int) -> Result<Option<Arc<T>>> {
@@ -235,7 +235,7 @@ where
     T: ToSql,
 {
     /// Initializes the aggregation context. Will be called prior to the first
-    /// call to `step()` to set up the context for an invocation of the
+    /// call to [`step()`](Aggregate::step) to set up the context for an invocation of the
     /// function. (Note: `init()` will not be called if there are no rows.)
     fn init(&self) -> A;
 
@@ -244,9 +244,9 @@ where
     fn step(&self, _: &mut Context<'_>, _: &mut A) -> Result<()>;
 
     /// Computes and returns the final result. Will be called exactly once for
-    /// each invocation of the function. If `step()` was called at least
+    /// each invocation of the function. If [`step()`](Aggregate::step) was called at least
     /// once, will be given `Some(A)` (the same `A` as was created by
-    /// `init` and given to `step`); if `step()` was not called (because
+    /// [`init`](Aggregate::init) and given to [`step`](Aggregate::step)); if [`step()`](Aggregate::step) was not called (because
     /// the function is running against 0 rows), will be given `None`.
     fn finalize(&self, _: Option<A>) -> Result<T>;
 }
@@ -309,7 +309,7 @@ impl Connection {
     /// given the same input, `deterministic` should be `true`.
     ///
     /// The function will remain available until the connection is closed or
-    /// until it is explicitly removed via `remove_function`.
+    /// until it is explicitly removed via [`remove_function`](Connection::remove_function).
     ///
     /// # Example
     ///
@@ -405,7 +405,7 @@ impl Connection {
     /// database connection.
     ///
     /// `fn_name` and `n_arg` should match the name and number of arguments
-    /// given to `create_scalar_function` or `create_aggregate_function`.
+    /// given to [`create_scalar_function`](Connection::create_scalar_function) or [`create_aggregate_function`](Connection::create_aggregate_function).
     ///
     /// # Failure
     ///
diff --git a/src/params.rs b/src/params.rs
index 21717fb..121d375 100644
--- a/src/params.rs
+++ b/src/params.rs
@@ -21,7 +21,7 @@ use sealed::Sealed;
 ///
 /// Many functions in this library let you pass parameters to SQLite. Doing this
 /// lets you avoid any risk of SQL injection, and is simpler than escaping
-/// things manually. Aside from deprecated functions and a few helpers, this is
+/// things manually. Aside from deprecated functions and a few helpers, this is
 /// indicated by the function taking a generic argument that implements `Params`
 /// (this trait).
 ///
diff --git a/src/row.rs b/src/row.rs
index 780916d..43fd89f 100644
--- a/src/row.rs
+++ b/src/row.rs
@@ -29,7 +29,7 @@ impl<'stmt> Rows<'stmt> {
     /// This interface is not compatible with Rust's `Iterator` trait, because
     /// the lifetime of the returned row is tied to the lifetime of `self`.
     /// This is a fallible "streaming iterator". For a more natural interface,
-    /// consider using `query_map` or `query_and_then` instead, which
+    /// consider using [`query_map`](crate::Statement::query_map) or [`query_and_then`](crate::Statement::query_and_then) instead, which
     /// return types that implement `Iterator`.
     #[allow(clippy::should_implement_trait)] // cannot implement Iterator
     #[inline]
diff --git a/src/session.rs b/src/session.rs
index 61cc285..b41106a 100644
--- a/src/session.rs
+++ b/src/session.rs
@@ -411,7 +411,7 @@ impl Drop for ChangesetIter<'_> {
 }
 
 /// `feature = "session"` An item passed to a conflict-handler by
-/// `Connection::apply`, or an item generated by `ChangesetIter::next`.
+/// [`Connection::apply`](crate::Connection::apply), or an item generated by [`ChangesetIter::next`](ChangesetIter::next).
 // TODO enum ? Delete, Insert, Update, ...
 pub struct ChangesetItem {
     it: *mut ffi::sqlite3_changeset_iter,
diff --git a/src/statement.rs b/src/statement.rs
index c7b2f52..b749e26 100644
--- a/src/statement.rs
+++ b/src/statement.rs
@@ -114,7 +114,7 @@ impl Statement<'_> {
     ///
     /// # Note
     ///
-    /// This function is a convenience wrapper around `execute()` intended for
+    /// This function is a convenience wrapper around [`execute()`](Statement::execute) intended for
     /// queries that insert a single item. It is possible to misuse this
     /// function in a way that it cannot detect, such as by calling it on a
     /// statement which _updates_ a single
@@ -136,8 +136,8 @@ impl Statement<'_> {
     /// rows.
     ///
     /// Due to lifetime restricts, the rows handle returned by `query` does not
-    /// implement the `Iterator` trait. Consider using `query_map` or
-    /// `query_and_then` instead, which do.
+    /// implement the `Iterator` trait. Consider using [`query_map`](Statement::query_map) or
+    /// [`query_and_then`](Statement::query_and_then) instead, which do.
     ///
     /// ## Example
     ///
@@ -437,7 +437,7 @@ impl Statement<'_> {
     /// ignored.
     ///
     /// Returns `Err(QueryReturnedNoRows)` if no results are returned. If the
-    /// query truly is optional, you can call `.optional()` on the result of
+    /// query truly is optional, you can call [`.optional()`](crate::OptionalExtension::optional) on the result of
     /// this to get a `Result<Option<T>>` (requires that the trait `rusqlite::OptionalExtension`
     /// is imported).
     ///
@@ -464,7 +464,7 @@ impl Statement<'_> {
     /// ignored.
     ///
     /// Returns `Err(QueryReturnedNoRows)` if no results are returned. If the
-    /// query truly is optional, you can call `.optional()` on the result of
+    /// query truly is optional, you can call [`.optional()`](crate::OptionalExtension::optional) on the result of
     /// this to get a `Result<Option<T>>` (requires that the trait `rusqlite::OptionalExtension`
     /// is imported).
     ///
@@ -493,7 +493,7 @@ impl Statement<'_> {
         self.finalize_()
     }
 
-    /// Return the (one-based) index of an SQL parameter given its name.
+    /// Return the (one-based) index of an SQL parameter given its name.
     ///
     /// Note that the initial ":" or "$" or "@" or "?" used to specify the
     /// parameter is included as part of the name.
@@ -563,7 +563,7 @@ impl Statement<'_> {
 
     /// Low level API to directly bind a parameter to a given index.
     ///
-    /// Note that the index is one-based, that is, the first parameter index is
+    /// Note that the index is one-based, that is, the first parameter index is
     /// 1 and not 0. This is consistent with the SQLite API and the values given
     /// to parameters bound as `?NNN`.
     ///
diff --git a/src/transaction.rs b/src/transaction.rs
index 0f00532..2192ce7 100644
--- a/src/transaction.rs
+++ b/src/transaction.rs
@@ -377,8 +377,8 @@ impl Connection {
     /// Begin a new transaction with the default behavior (DEFERRED).
     ///
     /// The transaction defaults to rolling back when it is dropped. If you
-    /// want the transaction to commit, you must call `commit` or
-    /// `set_drop_behavior(DropBehavior::Commit)`.
+    /// want the transaction to commit, you must call [`commit`](Transaction::commit) or
+    /// [`set_drop_behavior(DropBehavior::Commit)`](Transaction::set_drop_behavior).
     ///
     /// ## Example
     ///
@@ -406,7 +406,7 @@ impl Connection {
 
     /// Begin a new transaction with a specified behavior.
     ///
-    /// See `transaction`.
+    /// See [`transaction`](Connection::transaction).
     ///
     /// # Failure
     ///
@@ -457,8 +457,8 @@ impl Connection {
     /// Begin a new savepoint with the default behavior (DEFERRED).
     ///
     /// The savepoint defaults to rolling back when it is dropped. If you want
-    /// the savepoint to commit, you must call `commit` or
-    /// `set_drop_behavior(DropBehavior::Commit)`.
+    /// the savepoint to commit, you must call [`commit`](Savepoint::commit) or
+    /// [`set_drop_behavior(DropBehavior::Commit)`](Savepoint::set_drop_behavior).
     ///
     /// ## Example
     ///
@@ -486,7 +486,7 @@ impl Connection {
 
     /// Begin a new savepoint with a specified name.
     ///
-    /// See `savepoint`.
+    /// See [`savepoint`](Connection::savepoint).
     ///
     /// # Failure
     ///
diff --git a/src/types/from_sql.rs b/src/types/from_sql.rs
index cebffa8..3f0714c 100644
--- a/src/types/from_sql.rs
+++ b/src/types/from_sql.rs
@@ -3,7 +3,7 @@ use std::convert::TryInto;
 use std::error::Error;
 use std::fmt;
 
-/// Enum listing possible errors from `FromSql` trait.
+/// Enum listing possible errors from [`FromSql`] trait.
 #[derive(Debug)]
 #[non_exhaustive]
 pub enum FromSqlError {
@@ -26,7 +26,7 @@ pub enum FromSqlError {
     #[cfg(feature = "uuid")]
     InvalidUuidSize(usize),
 
-    /// An error case available for implementors of the `FromSql` trait.
+    /// An error case available for implementors of the [`FromSql`] trait.
     Other(Box<dyn Error + Send + Sync + 'static>),
 }
 
@@ -72,7 +72,7 @@ impl Error for FromSqlError {
     }
 }
 
-/// Result type for implementors of the `FromSql` trait.
+/// Result type for implementors of the [`FromSql`] trait.
 pub type FromSqlResult<T> = Result<T, FromSqlError>;
 
 /// A trait for types that can be created from a SQLite value.
diff --git a/src/types/mod.rs b/src/types/mod.rs
index a94747d..48d93ba 100644
--- a/src/types/mod.rs
+++ b/src/types/mod.rs
@@ -1,7 +1,7 @@
 //! Traits dealing with SQLite data types.
 //!
 //! SQLite uses a [dynamic type system](https://www.sqlite.org/datatype3.html). Implementations of
-//! the `ToSql` and `FromSql` traits are provided for the basic types that
+//! the [`ToSql`] and [`FromSql`] traits are provided for the basic types that
 //! SQLite provides methods for:
 //!
 //! * Strings (`String` and `&str`)
@@ -11,17 +11,17 @@
 //! The number situation is a little complicated due to the fact that all
 //! numbers in SQLite are stored as `INTEGER` (`i64`) or `REAL` (`f64`).
 //!
-//! `ToSql` and `FromSql` are implemented for all primitive number types.
-//! `FromSql` has different behaviour depending on the SQL and Rust types, and
+//! [`ToSql`] and [`FromSql`] are implemented for all primitive number types.
+//! [`FromSql`] has different behaviour depending on the SQL and Rust types, and
 //! the value.
 //!
-//! * `INTEGER` to integer: returns an `Error::IntegralValueOutOfRange` error if
+//! * `INTEGER` to integer: returns an [`Error::IntegralValueOutOfRange`](crate::Error::IntegralValueOutOfRange) error if
 //!   the value does not fit in the Rust type.
-//! * `REAL` to integer: always returns an `Error::InvalidColumnType` error.
+//! * `REAL` to integer: always returns an [`Error::InvalidColumnType`](crate::Error::InvalidColumnType) error.
 //! * `INTEGER` to float: casts using `as` operator. Never fails.
 //! * `REAL` to float: casts using `as` operator. Never fails.
 //!
-//! `ToSql` always succeeds except when storing a `u64` or `usize` value that
+//! [`ToSql`] always succeeds except when storing a `u64` or `usize` value that
 //! cannot fit in an `INTEGER` (`i64`). Also note that SQLite ignores column
 //! types, so if you store an `i64` in a column with type `REAL` it will be
 //! stored as an `INTEGER`, not a `REAL`.
@@ -61,8 +61,8 @@ impl ToSql for DateTimeSql {
 
 "##
 )]
-//! `ToSql` and `FromSql` are also implemented for `Option<T>` where `T`
-//! implements `ToSql` or `FromSql` for the cases where you want to know if a
+//! [`ToSql`] and [`FromSql`] are also implemented for `Option<T>` where `T`
+//! implements [`ToSql`] or [`FromSql`] for the cases where you want to know if a
 //! value was NULL (which gets translated to `None`).
 
 pub use self::from_sql::{FromSql, FromSqlError, FromSqlResult};
diff --git a/src/types/serde_json.rs b/src/types/serde_json.rs
index 1087144..f018032 100644
--- a/src/types/serde_json.rs
+++ b/src/types/serde_json.rs
@@ -1,4 +1,4 @@
-//! `ToSql` and `FromSql` implementation for JSON `Value`.
+//! [`ToSql`] and [`FromSql`] implementation for JSON `Value`.
 
 use serde_json::Value;
 
diff --git a/src/types/time.rs b/src/types/time.rs
index cac9d1c..14c92e1 100644
--- a/src/types/time.rs
+++ b/src/types/time.rs
@@ -1,4 +1,4 @@
-//! `ToSql` and `FromSql` implementation for [`time::OffsetDateTime`].
+//! [`ToSql`] and [`FromSql`] implementation for [`time::OffsetDateTime`].
 use crate::types::{FromSql, FromSqlError, FromSqlResult, ToSql, ToSqlOutput, ValueRef};
 use crate::Result;
 use time::{OffsetDateTime, PrimitiveDateTime, UtcOffset};
diff --git a/src/types/to_sql.rs b/src/types/to_sql.rs
index be88941..1bf7711 100644
--- a/src/types/to_sql.rs
+++ b/src/types/to_sql.rs
@@ -6,7 +6,7 @@ use std::borrow::Cow;
 use std::convert::TryFrom;
 
 /// `ToSqlOutput` represents the possible output types for implementers of the
-/// `ToSql` trait.
+/// [`ToSql`] trait.
 #[derive(Clone, Debug, PartialEq)]
 #[non_exhaustive]
 pub enum ToSqlOutput<'a> {
@@ -91,7 +91,7 @@ impl ToSql for ToSqlOutput<'_> {
 }
 
 /// A trait for types that can be converted into SQLite values. Returns
-/// `Error::ToSqlConversionFailure` if the conversion fails.
+/// [`Error::ToSqlConversionFailure`] if the conversion fails.
 pub trait ToSql {
     /// Converts Rust value to SQLite value
     fn to_sql(&self) -> Result<ToSqlOutput<'_>>;
diff --git a/src/types/url.rs b/src/types/url.rs
index a74604c..8991385 100644
--- a/src/types/url.rs
+++ b/src/types/url.rs
@@ -1,4 +1,4 @@
-//! `ToSql` and `FromSql` implementation for [`url::Url`].
+//! [`ToSql`] and [`FromSql`] implementation for [`url::Url`].
 use crate::types::{FromSql, FromSqlError, FromSqlResult, ToSql, ToSqlOutput, ValueRef};
 use crate::Result;
 use url::Url;
diff --git a/src/types/value.rs b/src/types/value.rs
index e460e82..bccbef6 100644
--- a/src/types/value.rs
+++ b/src/types/value.rs
@@ -3,7 +3,7 @@ use super::{Null, Type};
 /// Owning [dynamic type value](http://sqlite.org/datatype3.html). Value's type is typically
 /// dictated by SQLite (not by the caller).
 ///
-/// See [`ValueRef`](enum.ValueRef.html) for a non-owning dynamic type value.
+/// See [`ValueRef`](crate::types::ValueRef) for a non-owning dynamic type value.
 #[derive(Clone, Debug, PartialEq)]
 pub enum Value {
     /// The value is a `NULL` value.
diff --git a/src/types/value_ref.rs b/src/types/value_ref.rs
index 9c97a52..b95521b 100644
--- a/src/types/value_ref.rs
+++ b/src/types/value_ref.rs
@@ -4,7 +4,7 @@ use crate::types::{FromSqlError, FromSqlResult};
 /// A non-owning [dynamic type value](http://sqlite.org/datatype3.html). Typically the
 /// memory backing this value is owned by SQLite.
 ///
-/// See [`Value`](enum.Value.html) for an owning dynamic type value.
+/// See [`Value`](Value) for an owning dynamic type value.
 #[derive(Copy, Clone, Debug, PartialEq)]
 pub enum ValueRef<'a> {
     /// The value is a `NULL` value.
@@ -35,7 +35,7 @@ impl ValueRef<'_> {
 
 impl<'a> ValueRef<'a> {
     /// If `self` is case `Integer`, returns the integral value. Otherwise,
-    /// returns `Err(Error::InvalidColumnType)`.
+    /// returns [`Err(Error::InvalidColumnType)`](crate::Error::InvalidColumnType).
     #[inline]
     pub fn as_i64(&self) -> FromSqlResult<i64> {
         match *self {
@@ -45,7 +45,7 @@ impl<'a> ValueRef<'a> {
     }
 
     /// If `self` is case `Real`, returns the floating point value. Otherwise,
-    /// returns `Err(Error::InvalidColumnType)`.
+    /// returns [`Err(Error::InvalidColumnType)`](crate::Error::InvalidColumnType).
     #[inline]
     pub fn as_f64(&self) -> FromSqlResult<f64> {
         match *self {
@@ -55,7 +55,7 @@ impl<'a> ValueRef<'a> {
     }
 
     /// If `self` is case `Text`, returns the string value. Otherwise, returns
-    /// `Err(Error::InvalidColumnType)`.
+    /// [`Err(Error::InvalidColumnType)`](crate::Error::InvalidColumnType).
     #[inline]
     pub fn as_str(&self) -> FromSqlResult<&'a str> {
         match *self {
@@ -67,7 +67,7 @@ impl<'a> ValueRef<'a> {
     }
 
     /// If `self` is case `Blob`, returns the byte slice. Otherwise, returns
-    /// `Err(Error::InvalidColumnType)`.
+    /// [`Err(Error::InvalidColumnType)`](crate::Error::InvalidColumnType).
     #[inline]
     pub fn as_blob(&self) -> FromSqlResult<&'a [u8]> {
         match *self {
diff --git a/src/vtab/mod.rs b/src/vtab/mod.rs
index 76d323f..ee095f0 100644
--- a/src/vtab/mod.rs
+++ b/src/vtab/mod.rs
@@ -1,10 +1,10 @@
 //! `feature = "vtab"` Create virtual tables.
 //!
 //! Follow these steps to create your own virtual table:
-//! 1. Write implemenation of `VTab` and `VTabCursor` traits.
-//! 2. Create an instance of the `Module` structure specialized for `VTab` impl.
+//! 1. Write implemenation of [`VTab`] and [`VTabCursor`] traits.
+//! 2. Create an instance of the [`Module`] structure specialized for [`VTab`] impl.
 //! from step 1.
-//! 3. Register your `Module` structure using `Connection.create_module`.
+//! 3. Register your [`Module`] structure using [`Connection::create_module`].
 //! 4. Run a `CREATE VIRTUAL TABLE` command that specifies the new module in the
 //! `USING` clause.
 //!
@@ -205,7 +205,7 @@ impl VTabConnection {
 ///
 /// (See [SQLite doc](https://sqlite.org/c3ref/vtab.html))
 pub unsafe trait VTab<'vtab>: Sized {
-    /// Client data passed to `Connection::create_module`.
+    /// Client data passed to [`Connection::create_module`].
     type Aux;
     /// Specific cursor implementation
     type Cursor: VTabCursor;
@@ -237,7 +237,7 @@ pub trait CreateVTab<'vtab>: VTab<'vtab> {
     /// database connection that is executing the CREATE VIRTUAL TABLE
     /// statement.
     ///
-    /// Call `connect` by default.
+    /// Call [`connect`](VTab::connect) by default.
     /// (See [SQLite doc](https://sqlite.org/vtab.html#the_xcreate_method))
     fn create(
         db: &mut VTabConnection,
@@ -248,7 +248,7 @@ pub trait CreateVTab<'vtab>: VTab<'vtab> {
     }
 
     /// Destroy the underlying table implementation. This method undoes the work
-    /// of `create`.
+    /// of [`create`](CreateVTab::create).
     ///
     /// Do nothing by default.
     /// (See [SQLite doc](https://sqlite.org/vtab.html#the_xdestroy_method))
@@ -302,7 +302,7 @@ impl From<u8> for IndexConstraintOp {
 }
 
 /// `feature = "vtab"` Pass information into and receive the reply from the
-/// `VTab.best_index` method.
+/// [`VTab::best_index`] method.
 ///
 /// (See [SQLite doc](http://sqlite.org/c3ref/index_info.html))
 pub struct IndexInfo(*mut ffi::sqlite3_index_info);
@@ -334,7 +334,7 @@ impl IndexInfo {
         unsafe { (*self.0).nOrderBy as usize }
     }
 
-    /// Information about what parameters to pass to `VTabCursor.filter`.
+    /// Information about what parameters to pass to [`VTabCursor::filter`].
     #[inline]
     pub fn constraint_usage(&mut self, constraint_idx: usize) -> IndexConstraintUsage<'_> {
         let constraint_usages = unsafe {
@@ -425,11 +425,11 @@ impl IndexConstraint<'_> {
 }
 
 /// `feature = "vtab"` Information about what parameters to pass to
-/// `VTabCursor.filter`.
+/// [`VTabCursor::filter`].
 pub struct IndexConstraintUsage<'a>(&'a mut ffi::sqlite3_index_constraint_usage);
 
 impl IndexConstraintUsage<'_> {
-    /// if `argv_index` > 0, constraint is part of argv to `VTabCursor.filter`
+    /// if `argv_index` > 0, constraint is part of argv to [`VTabCursor::filter`]
     #[inline]
     pub fn set_argv_index(&mut self, argv_index: c_int) {
         self.0.argvIndex = argv_index;
@@ -495,7 +495,7 @@ pub unsafe trait VTabCursor: Sized {
     /// Begin a search of a virtual table.
     /// (See [SQLite doc](https://sqlite.org/vtab.html#the_xfilter_method))
     fn filter(&mut self, idx_num: c_int, idx_str: Option<&str>, args: &Values<'_>) -> Result<()>;
-    /// Advance cursor to the next row of a result set initiated by `filter`.
+    /// Advance cursor to the next row of a result set initiated by [`filter`](VTabCursor::filter).
     /// (See [SQLite doc](https://sqlite.org/vtab.html#the_xnext_method))
     fn next(&mut self) -> Result<()>;
     /// Must return `false` if the cursor currently points to a valid row of
@@ -512,7 +512,7 @@ pub unsafe trait VTabCursor: Sized {
     fn rowid(&self) -> Result<i64>;
 }
 
-/// `feature = "vtab"` Context is used by `VTabCursor.column` to specify the
+/// `feature = "vtab"` Context is used by [`VTabCursor::column`] to specify the
 /// cell value.
 pub struct Context(*mut ffi::sqlite3_context);
 
@@ -528,8 +528,8 @@ impl Context {
     // TODO sqlite3_vtab_nochange (http://sqlite.org/c3ref/vtab_nochange.html)
 }
 
-/// `feature = "vtab"` Wrapper to `VTabCursor.filter` arguments, the values
-/// requested by `VTab.best_index`.
+/// `feature = "vtab"` Wrapper to [`VTabCursor::filter`] arguments, the values
+/// requested by [`VTab::best_index`].
 pub struct Values<'a> {
     args: &'a [*mut ffi::sqlite3_value],
 }
@@ -606,7 +606,7 @@ impl<'a> IntoIterator for &'a Values<'a> {
     }
 }
 
-/// `Values` iterator.
+/// [`Values`] iterator.
 pub struct ValueIter<'a> {
     iter: slice::Iter<'a, *mut ffi::sqlite3_value>,
 }