scai 1.0.0 → 1.0.1

package/README.md

@@ -11,8 +11,13 @@
 - **No telemetry**
 - **Your code stays on your machine**
 
-Website: [https://scai.dk](https://scai.dk)  
-Built in Denmark / EU
+Built in Denmark and the EU with ❤️
+
+## Feedback And Support
+
+Reach out with feedback or questions:
+
+- Threads: [@scai.dk](https://threads.net/@scai.dk)
 
 ## What SCAI Is Good At
 
@@ -30,48 +35,6 @@ It is good at:
 - Helping with small light code changes in a limited area
 - Writing commit messages from staged changes
 
-## Current Limits
-
-SCAI is not a full replacement for a strong cloud coding agent.
-
-Today it works best when:
-
-- The task is about understanding code already in the repo
-- The task is about changing code that already exists
-- The change is small and local
-- You want grounded answers tied to real files
-
-Scai is not for:
-
-- Building brand new features from scratch
-- Large refactors across many files
-- Broad automated rewrite requests
-- Tasks that need bash actions, file creation, or other execution steps
-- Tasks that depend on web access or external services
-
-Important current limit:
-
-- SCAI can refactor code that already exists
-- SCAI is not yet a good fit for building new features end to end
-- SCAI does not currently act like a shell agent that runs bash steps to create files, wire systems together, or drive external tools
-- SCAI does not handle large file refactors well. It's capped at 1000 lines. 
-
-*SCAI was tested on an MacBook Pro M1 max (64GB). If you have a faster machine you may experience better performance and be able to use above 30b parameter models.*
-
-## Privacy And Cost
-
-SCAI runs with local models only.
-
-That means:
-
-- No API keys
-- No metered token billing
-- No remote prompt processing
-- No web browsing by the agent
-- No prompt injection from fetched web content
-
-This makes it a good fit for private code, internal repos, and teams that want a simple offline workflow.
-
 ## Getting Started
 
 Install the CLI:
@@ -99,7 +62,12 @@ This will:
 - Help you pick or install a local model
 - Start local indexing
 
-For most users, this is enough. Start with `scai setup` and only use the manual model commands below if you want to pick a specific model yourself.
+For most users, this is enough. Start with `scai setup` and only use the manual model commands below if you want to pick a specific model yourself. Run: 
+
+```bash
+scai status
+```
+...to check that setup worked.
 
 For best results on a 64 GB machine, use:
 
@@ -156,6 +124,48 @@ Help me add comments to this file.
 Help me make a small cleanup in this module.
 ```
 
+## Current Limits
+
+SCAI is not a full replacement for a strong cloud coding agent.
+
+Today it works best when:
+
+- The task is about understanding code already in the repo
+- The task is about changing code that already exists
+- The change is small and local
+- You want grounded answers tied to real files
+
+SCAI is not for:
+
+- Building brand new features from scratch
+- Large refactors across many files
+- Broad automated rewrite requests
+- Tasks that need bash actions, file creation, or other execution steps
+- Tasks that depend on web access or external services
+
+Important current limit:
+
+- SCAI can refactor code that already exists
+- SCAI is not yet a good fit for building new features end to end
+- SCAI does not currently act like a shell agent that runs bash steps to create files, wire systems together, or drive external tools
+- SCAI does not handle large file refactors well. It's capped at 1000 lines.
+
+*SCAI was tested on an MacBook Pro M1 max (64GB). If you have a faster machine you may experience better performance and be able to use above 30b parameter models.*
+
+## Privacy And Cost
+
+SCAI runs with local models only.
+
+That means:
+
+- No API keys
+- No metered token billing
+- No remote prompt processing
+- No web browsing by the agent
+- No prompt injection from fetched web content
+
+This makes it a good fit for private code, internal repos, and teams that want a simple offline workflow.
+
 ## Tasks And Resume
 
 SCAI can keep task context so you can continue later.
@@ -220,6 +230,14 @@ Only repositories that have been set up are available to the agent.
 
 SCAI uses a local Ollama server.
 
+Download Ollama here: [https://ollama.com/download](https://ollama.com/download)
+
+To download a model with Ollama, run:
+
+```bash
+ollama pull qwen3-coder:30b
+```
+
 Useful commands:
 
 ```bash
@@ -229,16 +247,10 @@ scai config show --raw
 scai config set-model qwen3-coder:30b
 ```
 
-If you want to manage models manually, download one with Ollama:
-
-```bash
-ollama pull gemma4:e4b-mlx
-```
-
-Then point SCAI at that model:
+Then point SCAI at the downloaded model:
 
 ```bash
-scai config set-model gemma4:e4b-mlx
+scai config set-model qwen3-coder:30b
 ```
 
 If `scai setup` worked, you usually do not need to touch this often.
@@ -271,3 +283,7 @@ If you need to inspect the current task state:
 scai context
 scai task show <id>
 ```
+
+## Maintainer Release Flow
+
+The CLI release workflow is documented in [release-procedure.md](/Users/rzs/dev/repos/scai/release-procedure.md).

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "scai",
-  "version": "1.0.0",
+  "version": "1.0.1",
   "type": "module",
   "bin": {
     "scai": "./dist/index.js"