update runner submodule

* feat(docs): README with Enchanted iOS app

* Update README.md

---------

Co-authored-by: Bruce MacDonald <brucewmacdonald@gmail.com>

README.md

@@ -57,6 +57,7 @@ Here are some example open-source models that can be downloaded:
 | Llama 2 70B        | 70B        | 39GB  | `ollama run llama2:70b`        |
 | Orca Mini          | 3B         | 1.9GB | `ollama run orca-mini`         |
 | Vicuna             | 7B         | 3.8GB | `ollama run vicuna`            |
+| LLaVA              | 7B         | 4.5GB | `ollama run llava`             |
 
 > Note: You should have at least 8 GB of RAM to run the 3B models, 16 GB to run the 7B models, and 32 GB to run the 13B models.
 
@@ -158,6 +159,13 @@ For multiline input, you can wrap text with `"""`:
 I'm a basic program that prints the famous "Hello, world!" message to the console.
 ```
 
+### Multimodal models
+
+```
+>>> What's in this image? /Users/jmorgan/Desktop/smile.png
+The image features a yellow smiley face, which is likely the central focus of the picture.
+```
+
 ### Pass in prompt as arguments
 
 ```
@@ -231,7 +239,7 @@ See the [API documentation](./docs/api.md) for all endpoints.
 ## Community Integrations
 
 ### Web & Desktop
-
+- [Bionic GPT](https://github.com/bionic-gpt/bionic-gpt)
 - [HTML UI](https://github.com/rtcfirefly/ollama-ui)
 - [Chatbot UI](https://github.com/ivanfioravanti/chatbot-ollama)
 - [Typescript UI](https://github.com/ollama-interface/Ollama-Gui?tab=readme-ov-file)
@@ -278,7 +286,8 @@ See the [API documentation](./docs/api.md) for all endpoints.
 
 ### Mobile
 
-- [Maid](https://github.com/danemadsen/Maid) (Mobile Artificial Intelligence Distribution)
+- [Enchanted](https://github.com/AugustDev/enchanted)
+- [Maid](https://github.com/danemadsen/Maid)
 
 ### Extensions & Plugins
 

api/types.go

@@ -59,13 +59,13 @@ type ChatRequest struct {
 type Message struct {
 	Role    string      `json:"role"` // one of ["system", "user", "assistant"]
 	Content string      `json:"content"`
-	Images  []ImageData `json:"images, omitempty"`
+	Images  []ImageData `json:"images,omitempty"`
 }
 
 type ChatResponse struct {
 	Model     string    `json:"model"`
 	CreatedAt time.Time `json:"created_at"`
-	Message   *Message  `json:"message,omitempty"`
+	Message   Message   `json:"message"`
 
 	Done bool `json:"done"`
 

cmd/cmd.go

@@ -19,7 +19,6 @@ import (
 	"path/filepath"
 	"regexp"
 	"runtime"
-	"slices"
 	"strings"
 	"syscall"
 	"time"
@@ -27,6 +26,7 @@ import (
 	"github.com/olekukonko/tablewriter"
 	"github.com/spf13/cobra"
 	"golang.org/x/crypto/ssh"
+	"golang.org/x/exp/slices"
 	"golang.org/x/term"
 
 	"github.com/jmorganca/ollama/api"
@@ -951,7 +951,8 @@ func generateInteractive(cmd *cobra.Command, opts generateOptions) error {
 					cmd.SetContext(ctx)
 				}
 				if len(opts.Images) == 0 {
-					fmt.Println("This model requires you to add a jpeg, png, or svg image.\n")
+					fmt.Println("This model requires you to add a jpeg, png, or svg image.")
+					fmt.Println()
 					prompt = ""
 					continue
 				}
@@ -1034,12 +1035,7 @@ func RunServer(cmd *cobra.Command, _ []string) error {
 		return err
 	}
 
-	var origins []string
-	if o := os.Getenv("OLLAMA_ORIGINS"); o != "" {
-		origins = strings.Split(o, ",")
-	}
-
-	return server.Serve(ln, origins)
+	return server.Serve(ln)
 }
 
 func getImageData(filePath string) ([]byte, error) {
@@ -1069,7 +1065,7 @@ func getImageData(filePath string) ([]byte, error) {
 	// Check if the file size exceeds 100MB
 	var maxSize int64 = 100 * 1024 * 1024 // 100MB in bytes
 	if info.Size() > maxSize {
-		return nil, fmt.Errorf("file size exceeds maximum limit (100MB).")
+		return nil, fmt.Errorf("file size exceeds maximum limit (100MB)")
 	}
 
 	buf = make([]byte, info.Size())

docs/api.md

@@ -39,6 +39,7 @@ Generate a response for a given prompt with a provided model. This is a streamin
 
 - `model`: (required) the [model name](#model-names)
 - `prompt`: the prompt to generate a response for
+- `images`: a list of base64-encoded images (for multimodal models such as `llava`)
 
 Advanced parameters (optional):
 
@@ -84,8 +85,6 @@ The final response in the stream also includes additional data about the generat
 
 - `total_duration`: time spent generating the response
 - `load_duration`: time spent in nanoseconds loading the model
-- `sample_count`: number of samples generated
-- `sample_duration`: time spent generating samples
 - `prompt_eval_count`: number of tokens in the prompt
 - `prompt_eval_duration`: time spent in nanoseconds evaluating the prompt
 - `eval_count`: number of tokens the response
@@ -104,8 +103,6 @@ To calculate how fast the response is generated in tokens per second (token/s),
   "done": true,
   "total_duration": 5589157167,
   "load_duration": 3013701500,
-  "sample_count": 114,
-  "sample_duration": 81442000,
   "prompt_eval_count": 46,
   "prompt_eval_duration": 1160282000,
   "eval_count": 113,
@@ -138,8 +135,6 @@ If `stream` is set to `false`, the response will be a single JSON object:
   "done": true,
   "total_duration": 5589157167,
   "load_duration": 3013701500,
-  "sample_count": 114,
-  "sample_duration": 81442000,
   "prompt_eval_count": 46,
   "prompt_eval_duration": 1160282000,
   "eval_count": 13,
@@ -147,6 +142,37 @@ If `stream` is set to `false`, the response will be a single JSON object:
 }
 ```
 
+#### Request (with images)
+
+To submit images to multimodal models such as `llava` or `bakllava`, provide a list of base64-encoded `images`:
+
+```shell
+curl http://localhost:11434/api/generate -d '{
+  "model": "llava",
+  "prompt":"What is in this picture?",
+  "stream": false,
+  "images": ["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"]
+}'
+```
+
+#### Response
+
+```
+{
+  "model": "llava",
+  "created_at": "2023-11-03T15:36:02.583064Z",
+  "response": "A happy cartoon character, which is cute and cheerful.",
+  "context": [1, 2, 3],
+  "done": true,
+  "total_duration": 14648695333,
+  "load_duration": 3302671417,
+  "prompt_eval_count": 14,
+  "prompt_eval_duration": 286243000,
+  "eval_count": 129,
+  "eval_duration": 10931424000
+}
+```
+
 #### Request (Raw Mode)
 
 In some cases you may wish to bypass the templating system and provide a full prompt. In this case, you can use the `raw` parameter to disable formatting.
@@ -282,8 +308,6 @@ curl http://localhost:11434/api/generate -d '{
   "done": true,
   "total_duration": 5589157167,
   "load_duration": 3013701500,
-  "sample_count": 114,
-  "sample_duration": 81442000,
   "prompt_eval_count": 46,
   "prompt_eval_duration": 1160282000,
   "eval_count": 13,
@@ -304,6 +328,12 @@ Generate the next message in a chat with a provided model. This is a streaming e
 - `model`: (required) the [model name](#model-names)
 - `messages`: the messages of the chat, this can be used to keep a chat memory
 
+The `message` object has the following fields:
+
+- `role`: the role of the message, either `system`, `user` or `assistant`
+- `content`: the content of the message
+- `images` (optional): a list of images to include in the message (for multimodal models such as `llava`)
+
 Advanced parameters (optional):
 
 - `format`: the format to return a response in. Currently the only accepted value is `json`
@@ -354,8 +384,6 @@ Final response:
   "done": true,
   "total_duration": 5589157167,
   "load_duration": 3013701500,
-  "sample_count": 114,
-  "sample_duration": 81442000,
   "prompt_eval_count": 46,
   "prompt_eval_duration": 1160282000,
   "eval_count": 113,
@@ -412,8 +440,6 @@ Final response:
   "done": true,
   "total_duration": 5589157167,
   "load_duration": 3013701500,
-  "sample_count": 114,
-  "sample_duration": 81442000,
   "prompt_eval_count": 46,
   "prompt_eval_duration": 1160282000,
   "eval_count": 113,
@@ -421,6 +447,23 @@ Final response:
 }
 ```
 
+#### Request (with images)
+
+Send a chat message with a conversation history.
+
+```shell
+curl http://localhost:11434/api/chat -d '{
+  "model": "llama2",
+  "messages": [
+    {
+      "role": "user",
+      "content": "what is in this image?",
+      "images": ["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"]
+    },
+  ]
+}'
+```
+
 ## Create a Model
 
 ```shell
@@ -548,7 +591,7 @@ A single JSON object will be returned.
 POST /api/show
 ```
 
-Show details about a model including modelfile, template, parameters, license, and system message.
+Show information about a model including details, modelfile, template, parameters, license, and system prompt.
 
 ### Parameters
 
@@ -568,10 +611,16 @@ curl http://localhost:11434/api/show -d '{
 
 ```json
 {
-  "license": "<contents of license block>",
-  "modelfile": "# Modelfile generated by \"ollama show\"\n# To build a new Modelfile based on this one, replace the FROM line with:\n# FROM llama2:latest\n\nFROM /Users/username/.ollama/models/blobs/sha256:8daa9615cce30c259a9555b1cc250d461d1bc69980a274b44d7eda0be78076d8\nTEMPLATE \"\"\"[INST] {{ if and .First .System }}<<SYS>>{{ .System }}<</SYS>>\n\n{{ end }}{{ .Prompt }} [/INST] \"\"\"\nSYSTEM \"\"\"\"\"\"\nPARAMETER stop [INST]\nPARAMETER stop [/INST]\nPARAMETER stop <<SYS>>\nPARAMETER stop <</SYS>>\n",
-  "parameters": "stop                           [INST]\nstop                           [/INST]\nstop                           <<SYS>>\nstop                           <</SYS>>",
-  "template": "[INST] {{ if and .First .System }}<<SYS>>{{ .System }}<</SYS>>\n\n{{ end }}{{ .Prompt }} [/INST] "
+  "modelfile": "# Modelfile generated by \"ollama show\"\n# To build a new Modelfile based on this one, replace the FROM line with:\n# FROM llava:latest\n\nFROM mike/llava:latest\nTEMPLATE \"\"\"\nUSER:{{ .Prompt }}\nASSISTANT:\n\"\"\"\nPARAMETER num_ctx 4096",
+  "parameters": "num_ctx                        4096",
+  "template": "\nUSER:{{ .Prompt }}\nASSISTANT:\n",
+  "license:": "<license>",
+  "details": {
+    "format": "gguf",
+    "families": ["llama", "clip"],
+    "parameter_size": "7B",
+    "quantization_level": "Q4_0"
+  }
 }
 ```
 

go.mod

@@ -7,11 +7,14 @@ require (
 	github.com/gin-gonic/gin v1.9.1
 	github.com/olekukonko/tablewriter v0.0.5
 	github.com/spf13/cobra v1.7.0
+	github.com/stretchr/testify v1.8.3
 	golang.org/x/sync v0.3.0
 )
 
 require (
+	github.com/davecgh/go-spew v1.1.1 // indirect
 	github.com/mattn/go-runewidth v0.0.14 // indirect
+	github.com/pmezard/go-difflib v1.0.0 // indirect
 	github.com/rivo/uniseg v0.2.0 // indirect
 )
 

llm/llama.go

@@ -548,17 +548,12 @@ const maxBufferSize = 512 * format.KiloByte
 const maxRetries = 6
 
 type PredictOpts struct {
-	Prompt           string
-	Format           string
-	Images           []api.ImageData
-	CheckpointStart  time.Time
-	CheckpointLoaded time.Time
+	Prompt string
+	Format string
+	Images []api.ImageData
 }
 
 type PredictResult struct {
-	CreatedAt          time.Time
-	TotalDuration      time.Duration
-	LoadDuration       time.Duration
 	Content            string
 	Done               bool
 	PromptEvalCount    int
@@ -681,16 +676,12 @@ func (llm *llama) Predict(ctx context.Context, predict PredictOpts, fn func(Pred
 
 				if p.Content != "" {
 					fn(PredictResult{
-						CreatedAt: time.Now().UTC(),
-						Content:   p.Content,
+						Content: p.Content,
 					})
 				}
 
 				if p.Stop {
 					fn(PredictResult{
-						CreatedAt:     time.Now().UTC(),
-						TotalDuration: time.Since(predict.CheckpointStart),
-
 						Done:               true,
 						PromptEvalCount:    p.Timings.PromptN,
 						PromptEvalDuration: parseDurationMs(p.Timings.PromptMS),

server/images_test.go

@@ -81,7 +81,7 @@ func TestChat(t *testing.T) {
 			Template: tt.template,
 		}
 		t.Run(tt.name, func(t *testing.T) {
-			got, err := m.ChatPrompt(tt.msgs)
+			got, _, err := m.ChatPrompt(tt.msgs)
 			if tt.wantErr != "" {
 				if err == nil {
 					t.Errorf("ChatPrompt() expected error, got nil")

server/routes.go

@@ -32,6 +32,10 @@ import (
 
 var mode string = gin.DebugMode
 
+type Server struct {
+	WorkDir string
+}
+
 func init() {
 	switch mode {
 	case gin.DebugMode:
@@ -261,12 +265,10 @@ func GenerateHandler(c *gin.Context) {
 
 			resp := api.GenerateResponse{
 				Model:     req.Model,
-				CreatedAt: r.CreatedAt,
+				CreatedAt: time.Now().UTC(),
 				Done:      r.Done,
 				Response:  r.Content,
 				Metrics: api.Metrics{
-					TotalDuration:      r.TotalDuration,
-					LoadDuration:       r.LoadDuration,
 					PromptEvalCount:    r.PromptEvalCount,
 					PromptEvalDuration: r.PromptEvalDuration,
 					EvalCount:          r.EvalCount,
@@ -274,13 +276,18 @@ func GenerateHandler(c *gin.Context) {
 				},
 			}
 
-			if r.Done && !req.Raw {
-				embd, err := loaded.runner.Encode(c.Request.Context(), prompt+generated.String())
-				if err != nil {
-					ch <- gin.H{"error": err.Error()}
-					return
+			if r.Done {
+				resp.TotalDuration = time.Since(checkpointStart)
+				resp.LoadDuration = checkpointLoaded.Sub(checkpointStart)
+
+				if !req.Raw {
+					embd, err := loaded.runner.Encode(c.Request.Context(), prompt+generated.String())
+					if err != nil {
+						ch <- gin.H{"error": err.Error()}
+						return
+					}
+					resp.Context = embd
 				}
-				resp.Context = embd
 			}
 
 			ch <- resp
@@ -288,11 +295,9 @@ func GenerateHandler(c *gin.Context) {
 
 		// Start prediction
 		predictReq := llm.PredictOpts{
-			Prompt:           prompt,
-			Format:           req.Format,
-			CheckpointStart:  checkpointStart,
-			CheckpointLoaded: checkpointLoaded,
-			Images:           req.Images,
+			Prompt: prompt,
+			Format: req.Format,
+			Images: req.Images,
 		}
 		if err := loaded.runner.Predict(c.Request.Context(), predictReq, fn); err != nil {
 			ch <- gin.H{"error": err.Error()}
@@ -799,27 +804,27 @@ var defaultAllowOrigins = []string{
 	"0.0.0.0",
 }
 
-func Serve(ln net.Listener, allowOrigins []string) error {
-	if noprune := os.Getenv("OLLAMA_NOPRUNE"); noprune == "" {
-		// clean up unused layers and manifests
-		if err := PruneLayers(); err != nil {
-			return err
-		}
+func NewServer() (*Server, error) {
+	workDir, err := os.MkdirTemp("", "ollama")
+	if err != nil {
+		return nil, err
+	}
 
-		manifestsPath, err := GetManifestPath()
-		if err != nil {
-			return err
-		}
+	return &Server{
+		WorkDir: workDir,
+	}, nil
+}
 
-		if err := PruneDirectory(manifestsPath); err != nil {
-			return err
-		}
+func (s *Server) GenerateRoutes() http.Handler {
+	var origins []string
+	if o := os.Getenv("OLLAMA_ORIGINS"); o != "" {
+		origins = strings.Split(o, ",")
 	}
 
 	config := cors.DefaultConfig()
 	config.AllowWildcard = true
 
-	config.AllowOrigins = allowOrigins
+	config.AllowOrigins = origins
 	for _, allowOrigin := range defaultAllowOrigins {
 		config.AllowOrigins = append(config.AllowOrigins,
 			fmt.Sprintf("http://%s", allowOrigin),
@@ -829,17 +834,11 @@ func Serve(ln net.Listener, allowOrigins []string) error {
 		)
 	}
 
-	workDir, err := os.MkdirTemp("", "ollama")
-	if err != nil {
-		return err
-	}
-	defer os.RemoveAll(workDir)
-
 	r := gin.Default()
 	r.Use(
 		cors.New(config),
 		func(c *gin.Context) {
-			c.Set("workDir", workDir)
+			c.Set("workDir", s.WorkDir)
 			c.Next()
 		},
 	)
@@ -867,8 +866,34 @@ func Serve(ln net.Listener, allowOrigins []string) error {
 		})
 	}
 
+	return r
+}
+
+func Serve(ln net.Listener) error {
+	if noprune := os.Getenv("OLLAMA_NOPRUNE"); noprune == "" {
+		// clean up unused layers and manifests
+		if err := PruneLayers(); err != nil {
+			return err
+		}
+
+		manifestsPath, err := GetManifestPath()
+		if err != nil {
+			return err
+		}
+
+		if err := PruneDirectory(manifestsPath); err != nil {
+			return err
+		}
+	}
+
+	s, err := NewServer()
+	if err != nil {
+		return err
+	}
+	r := s.GenerateRoutes()
+
 	log.Printf("Listening on %s (version %s)", ln.Addr(), version.Version)
-	s := &http.Server{
+	srvr := &http.Server{
 		Handler: r,
 	}
 
@@ -880,7 +905,7 @@ func Serve(ln net.Listener, allowOrigins []string) error {
 		if loaded.runner != nil {
 			loaded.runner.Close()
 		}
-		os.RemoveAll(workDir)
+		os.RemoveAll(s.WorkDir)
 		os.Exit(0)
 	}()
 
@@ -891,7 +916,7 @@ func Serve(ln net.Listener, allowOrigins []string) error {
 		}
 	}
 
-	return s.Serve(ln)
+	return srvr.Serve(ln)
 }
 
 func waitForStream(c *gin.Context, ch chan interface{}) {
@@ -988,7 +1013,7 @@ func ChatHandler(c *gin.Context) {
 
 	// an empty request loads the model
 	if len(req.Messages) == 0 {
-		c.JSON(http.StatusOK, api.ChatResponse{CreatedAt: time.Now().UTC(), Model: req.Model, Done: true})
+		c.JSON(http.StatusOK, api.ChatResponse{CreatedAt: time.Now().UTC(), Model: req.Model, Done: true, Message: api.Message{Role: "assistant"}})
 		return
 	}
 
@@ -1012,11 +1037,10 @@ func ChatHandler(c *gin.Context) {
 
 			resp := api.ChatResponse{
 				Model:     req.Model,
-				CreatedAt: r.CreatedAt,
+				CreatedAt: time.Now().UTC(),
+				Message:   api.Message{Role: "assistant", Content: r.Content},
 				Done:      r.Done,
 				Metrics: api.Metrics{
-					TotalDuration:      r.TotalDuration,
-					LoadDuration:       r.LoadDuration,
 					PromptEvalCount:    r.PromptEvalCount,
 					PromptEvalDuration: r.PromptEvalDuration,
 					EvalCount:          r.EvalCount,
@@ -1024,8 +1048,9 @@ func ChatHandler(c *gin.Context) {
 				},
 			}
 
-			if !r.Done {
-				resp.Message = &api.Message{Role: "assistant", Content: r.Content}
+			if r.Done {
+				resp.TotalDuration = time.Since(checkpointStart)
+				resp.LoadDuration = checkpointLoaded.Sub(checkpointStart)
 			}
 
 			ch <- resp
@@ -1033,11 +1058,9 @@ func ChatHandler(c *gin.Context) {
 
 		// Start prediction
 		predictReq := llm.PredictOpts{
-			Prompt:           prompt,
-			Format:           req.Format,
-			CheckpointStart:  checkpointStart,
-			CheckpointLoaded: checkpointLoaded,
-			Images:           images,
+			Prompt: prompt,
+			Format: req.Format,
+			Images: images,
 		}
 		if err := loaded.runner.Predict(c.Request.Context(), predictReq, fn); err != nil {
 			ch <- gin.H{"error": err.Error()}
@@ -1051,10 +1074,7 @@ func ChatHandler(c *gin.Context) {
 		for resp := range ch {
 			switch r := resp.(type) {
 			case api.ChatResponse:
-				if r.Message != nil {
-					sb.WriteString(r.Message.Content)
-				}
-
+				sb.WriteString(r.Message.Content)
 				final = r
 			case gin.H:
 				if errorMsg, ok := r["error"].(string); ok {
@@ -1070,7 +1090,7 @@ func ChatHandler(c *gin.Context) {
 			}
 		}
 
-		final.Message = &api.Message{Role: "assistant", Content: sb.String()}
+		final.Message = api.Message{Role: "assistant", Content: sb.String()}
 		c.JSON(http.StatusOK, final)
 		return
 	}

server/routes_test.go

@@ -0,0 +1,204 @@
+package server
+
+import (
+	"bytes"
+	"context"
+	"encoding/json"
+	"fmt"
+	"io"
+	"net/http"
+	"net/http/httptest"
+	"os"
+	"strings"
+	"testing"
+
+	"github.com/stretchr/testify/assert"
+
+	"github.com/jmorganca/ollama/api"
+	"github.com/jmorganca/ollama/parser"
+)
+
+func setupServer(t *testing.T) (*Server, error) {
+	t.Helper()
+
+	return NewServer()
+}
+
+func Test_Routes(t *testing.T) {
+	type testCase struct {
+		Name     string
+		Method   string
+		Path     string
+		Setup    func(t *testing.T, req *http.Request)
+		Expected func(t *testing.T, resp *http.Response)
+	}
+
+	createTestFile := func(t *testing.T, name string) string {
+		f, err := os.CreateTemp(t.TempDir(), name)
+		assert.Nil(t, err)
+		defer f.Close()
+
+		_, err = f.Write([]byte("GGUF"))
+		assert.Nil(t, err)
+		_, err = f.Write([]byte{0x2, 0})
+		assert.Nil(t, err)
+
+		return f.Name()
+	}
+
+	createTestModel := func(t *testing.T, name string) {
+		fname := createTestFile(t, "ollama-model")
+
+		modelfile := strings.NewReader(fmt.Sprintf("FROM %s", fname))
+		commands, err := parser.Parse(modelfile)
+		assert.Nil(t, err)
+		fn := func(resp api.ProgressResponse) {
+			t.Logf("Status: %s", resp.Status)
+		}
+		err = CreateModel(context.TODO(), name, "", commands, fn)
+		assert.Nil(t, err)
+	}
+
+	testCases := []testCase{
+		{
+			Name:   "Version Handler",
+			Method: http.MethodGet,
+			Path:   "/api/version",
+			Setup: func(t *testing.T, req *http.Request) {
+			},
+			Expected: func(t *testing.T, resp *http.Response) {
+				contentType := resp.Header.Get("Content-Type")
+				assert.Equal(t, contentType, "application/json; charset=utf-8")
+				body, err := io.ReadAll(resp.Body)
+				assert.Nil(t, err)
+				assert.Equal(t, `{"version":"0.0.0"}`, string(body))
+			},
+		},
+		{
+			Name:   "Tags Handler (no tags)",
+			Method: http.MethodGet,
+			Path:   "/api/tags",
+			Expected: func(t *testing.T, resp *http.Response) {
+				contentType := resp.Header.Get("Content-Type")
+				assert.Equal(t, contentType, "application/json; charset=utf-8")
+				body, err := io.ReadAll(resp.Body)
+				assert.Nil(t, err)
+
+				var modelList api.ListResponse
+
+				err = json.Unmarshal(body, &modelList)
+				assert.Nil(t, err)
+
+				assert.Equal(t, 0, len(modelList.Models))
+			},
+		},
+		{
+			Name:   "Tags Handler (yes tags)",
+			Method: http.MethodGet,
+			Path:   "/api/tags",
+			Setup: func(t *testing.T, req *http.Request) {
+				createTestModel(t, "test-model")
+			},
+			Expected: func(t *testing.T, resp *http.Response) {
+				contentType := resp.Header.Get("Content-Type")
+				assert.Equal(t, contentType, "application/json; charset=utf-8")
+				body, err := io.ReadAll(resp.Body)
+				assert.Nil(t, err)
+
+				var modelList api.ListResponse
+				err = json.Unmarshal(body, &modelList)
+				assert.Nil(t, err)
+
+				assert.Equal(t, 1, len(modelList.Models))
+				assert.Equal(t, modelList.Models[0].Name, "test-model:latest")
+			},
+		},
+		{
+			Name:   "Create Model Handler",
+			Method: http.MethodPost,
+			Path:   "/api/create",
+			Setup: func(t *testing.T, req *http.Request) {
+				f, err := os.CreateTemp(t.TempDir(), "ollama-model")
+				assert.Nil(t, err)
+				defer f.Close()
+
+				stream := false
+				createReq := api.CreateRequest{
+					Name:      "t-bone",
+					Modelfile: fmt.Sprintf("FROM %s", f.Name()),
+					Stream:    &stream,
+				}
+				jsonData, err := json.Marshal(createReq)
+				assert.Nil(t, err)
+
+				req.Body = io.NopCloser(bytes.NewReader(jsonData))
+			},
+			Expected: func(t *testing.T, resp *http.Response) {
+				contentType := resp.Header.Get("Content-Type")
+				assert.Equal(t, "application/json", contentType)
+				_, err := io.ReadAll(resp.Body)
+				assert.Nil(t, err)
+				assert.Equal(t, resp.StatusCode, 200)
+
+				model, err := GetModel("t-bone")
+				assert.Nil(t, err)
+				assert.Equal(t, "t-bone:latest", model.ShortName)
+			},
+		},
+		{
+			Name:   "Copy Model Handler",
+			Method: http.MethodPost,
+			Path:   "/api/copy",
+			Setup: func(t *testing.T, req *http.Request) {
+				createTestModel(t, "hamshank")
+				copyReq := api.CopyRequest{
+					Source:      "hamshank",
+					Destination: "beefsteak",
+				}
+				jsonData, err := json.Marshal(copyReq)
+				assert.Nil(t, err)
+
+				req.Body = io.NopCloser(bytes.NewReader(jsonData))
+			},
+			Expected: func(t *testing.T, resp *http.Response) {
+				model, err := GetModel("beefsteak")
+				assert.Nil(t, err)
+				assert.Equal(t, "beefsteak:latest", model.ShortName)
+			},
+		},
+	}
+
+	s, err := setupServer(t)
+	assert.Nil(t, err)
+
+	router := s.GenerateRoutes()
+
+	httpSrv := httptest.NewServer(router)
+	t.Cleanup(httpSrv.Close)
+
+	workDir, err := os.MkdirTemp("", "ollama-test")
+	assert.Nil(t, err)
+	defer os.RemoveAll(workDir)
+	os.Setenv("OLLAMA_MODELS", workDir)
+
+	for _, tc := range testCases {
+		t.Logf("Running Test: [%s]", tc.Name)
+		u := httpSrv.URL + tc.Path
+		req, err := http.NewRequestWithContext(context.TODO(), tc.Method, u, nil)
+		assert.Nil(t, err)
+
+		if tc.Setup != nil {
+			tc.Setup(t, req)
+		}
+
+		resp, err := httpSrv.Client().Do(req)
+		defer resp.Body.Close()
+		assert.Nil(t, err)
+
+		if tc.Expected != nil {
+			tc.Expected(t, resp)
+		}
+
+	}
+
+}