wip: remove opencode, searxng, resources; fix user settings read

Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
2026-03-06 07:27:19 +00:00
parent 5925ac49a1
commit 7bbcccabf1
46 changed files with 325 additions and 2037 deletions
@@ -1,51 +0,0 @@
-# Resource Configuration Guide
-
-You are an AI assistant helping users configure **resources** — external services and APIs that Officer connects to.
-
-## What is a Resource?
-
-A resource represents an external service (e.g., a TTS server, an STT API, an OCR endpoint). Each resource has:
- A `RESOURCE.md` file with frontmatter metadata (name, description) and a markdown body
- A `config.json` file with flat key-value connection settings (all strings)
-
-## config.json Format
-
-The config is a flat JSON object where every value is a string. Empty string means "not set".
-
-Base fields (always present):
- `url` — the service endpoint URL
- `api_key` — API key or token for authentication
- `username` — username for basic auth
- `password` — password for basic auth
-
-Additional fields vary per resource (e.g., `provider`, `model`, `voice`).
-
-Example:
-```json
-{
-  "url": "http://localhost:8000",
-  "api_key": "",
-  "username": "",
-  "password": "",
-  "provider": "openai",
-  "model": "tts-1",
-  "voice": "alloy"
-}
-```
-
-## How to Help the User
-
-1. **Ask what service they're connecting to** — provider name, URL, auth method
-2. **Fill in the config.json** — write the file with the values they provide
-3. **Explain each field** — tell them what each key does and what format it expects
-4. **For new resources**, also write the `RESOURCE.md` with an appropriate name and description in the frontmatter
-
-## RESOURCE.md Format
-
-```markdown
---
-name: Human Readable Name
-description: One-line description of what this resource does
---
-Optional longer markdown body with usage notes, compatible providers, etc.
-```
@@ -1,6 +0,0 @@
-# Optical Character Recognition — OCR API
-
- **Type:** OCR
- **Port:** 8082
- **Description:** HTTP server for optical character recognition. Compatible with OpenAI-style vision APIs. Configure the URL and model in the connection settings below.
- **Verify:** `curl -sf {url}/v1/models`
@@ -1,6 +0,0 @@
-# Speech to Text — Transcription API
-
- **Type:** STT
- **Port:** 8080
- **Description:** HTTP server for speech-to-text transcription. Compatible with the whisper.cpp server API. Configure the URL in the connection settings below.
- **Verify:** `curl -sf {url}/health`
@@ -1,6 +0,0 @@
-# Text to Speech — Synthesis API
-
- **Type:** TTS
- **Port:** 8000
- **Description:** HTTP server for text-to-speech synthesis. Compatible with the OpenAI audio/speech API (mlx-audio, Kokoro, etc.) and ElevenLabs. Configure the URL and credentials in the connection settings below.
- **Verify:** `curl -sf {url}/v1/models`
@@ -1,5 +0,0 @@
---
-name: Optical Character Recognition
-description: HTTP server for optical character recognition
---
-Compatible with OpenAI-style vision APIs for extracting text from images.
@@ -1,7 +0,0 @@
-{
-  "url": "",
-  "api_key": "",
-  "username": "",
-  "password": "",
-  "model": ""
-}
@@ -1,5 +0,0 @@
---
-name: Speech to Text
-description: HTTP server for speech-to-text transcription
---
-Compatible with the whisper.cpp server API and OpenAI-compatible transcription endpoints.
@@ -1,6 +0,0 @@
-{
-  "url": "",
-  "api_key": "",
-  "username": "",
-  "password": ""
-}
@@ -1,5 +0,0 @@
---
-name: Text to Speech
-description: HTTP server for text-to-speech synthesis
---
-Compatible with OpenAI audio/speech API (mlx-audio, Kokoro, etc.) and ElevenLabs.
@@ -1,9 +0,0 @@
-{
-  "url": "",
-  "api_key": "",
-  "username": "",
-  "password": "",
-  "provider": "",
-  "model": "",
-  "voice": ""
-}