diff --git a/Llama.hs b/Llama.hs
--- a/Llama.hs
+++ b/Llama.hs
@@ -1,4 +1,4 @@
-{-# LANGUAGE OverloadedStrings, DeriveGeneric, DuplicateRecordFields #-}
+{-# LANGUAGE OverloadedStrings, DeriveGeneric, DuplicateRecordFields, DataKinds, DerivingVia #-}
 
 module Llama where
 
@@ -8,7 +8,7 @@
 import Data.Default
 import Data.Text (Text)
 import Data.Word
-import GHC.Generics
+import Deriving.Aeson
 import Network.HTTP.Conduit
 import Network.HTTP.Simple hiding (httpLbs)
 import Network.HTTP.Types.Status
@@ -61,6 +61,23 @@
   } deriving (Show, Generic)
 instance FromJSON LlamaDetokenizeResponse
 
+-- |One model entry of the `/models` response
+data LlamaModelInfo = LlamaModelInfo
+  { lmiId :: Text
+  , lmiAliases :: [Text]
+  , lmiTags :: [Text]
+  , lmiObject :: Text
+  , lmiOwnedBy :: Text
+  , lmiCreated :: Word
+  } deriving (Show, Generic)
+  deriving (FromJSON) via CustomJSON '[FieldLabelModifier '[StripPrefix "lmi", CamelToSnake]] LlamaModelInfo
+
+-- |The whole `/models` response
+newtype LlamaModelsResponse = LlamaModelsResponse
+  { lmrData :: [LlamaModelInfo]
+  } deriving (Show, Generic)
+  deriving (FromJSON) via CustomJSON '[FieldLabelModifier '[StripPrefix "lmr", CamelToSnake]] LlamaModelsResponse
+
 data Health = HealthOk | HealthNok deriving (Show)
 
 -- llama.cpp rejects requests with null options since https://github.com/ggml-org/llama.cpp/pull/24150
@@ -230,3 +247,11 @@
   pure $ if responseStatus response == ok200
        then HealthOk
        else HealthNok
+
+-- |Returns the models served by the llama-server
+models :: URL -> IO (Maybe [LlamaModelInfo])
+models url = do
+  let request = parseRequest_ $ url ++ "/models"
+  response <- httpLBS request
+  decoded <- llamaDecode (responseBody response)
+  return $ decoded >>= (\(LlamaModelsResponse result) -> Just result)
diff --git a/llama-cpp-haskell.cabal b/llama-cpp-haskell.cabal
--- a/llama-cpp-haskell.cabal
+++ b/llama-cpp-haskell.cabal
@@ -1,6 +1,6 @@
 cabal-version:      2.2
 name:               llama-cpp-haskell
-version:            0.3.0.2
+version:            0.3.1
 synopsis:           Haskell bindings for the llama.cpp llama-server and a simple CLI
 description:        This is the interface that allows one to interface with llama-server RPC API using Haskell concepts. It also includes a `llamacall` binary to do it from your favorite command line shell and use it in scripting.
 license:            AGPL-3.0-only
@@ -20,7 +20,7 @@
 Source-repository this
   type:              git
   location:          https://github.com/l29ah/llama-cpp-haskell.git
-  tag:               0.3.0.2
+  tag:               0.3.1
 
 common stuff
     ghc-options: -Wall
@@ -36,6 +36,7 @@
                     , http-types ^>= 0.12
                     , bytestring ^>= 0.12.2.0
                     , attoparsec ^>= 0.14.4
+                    , deriving-aeson ^>= 0.2.9
                     , data-default >= 0.7 && < 0.9
 
 library