> For the complete documentation index, see [llms.txt](https://yvechat.gitbook.io/wp/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://yvechat.gitbook.io/wp/capabilities.md).
# Capabilities
Every capability carries a status. Nothing appears here without one.
| Status | Meaning |
| ---------------- | ----------------------------------------------------------------------------------------------------------------------------------- |
| **Available** | Implemented and reachable by a user today |
| **Experimental** | Implemented and requiring manual configuration, or implemented at one layer and pending integration at another. Behavior may change |
| **Planned** | A stated intention with no implementation |
Every capability carries one of three labels
***
### Available
| Capability | Detail |
| ----------------------------- | ---------------------------------------------------------------------------------------------------------------------------------- |
| **Local language models** | Ollama-served models running on the user's hardware. No credential, no provider, no outbound request |
| **Dynamic model discovery** | The model selector is populated from live YveChat Engine discovery on every refresh. Newly installed models appear after a refresh |
| **Validated model downloads** | Explicit, validated Ollama pulls performed by YveChat Engine after user confirmation |
| **Local image generation** | ComfyUI, presented through a capability-aware image registry, in Image Studio |
| **Web search with citations** | Provider-neutral gateway returning cited results. Ollama Web Search and self-hosted SearXNG adapters are supported |
| **Conversation storage** | A SQLite database managed by YveChat Engine on the user's computer |
| **Local accounts** | Profiles held locally, with scrypt-hashed passwords of at least 12 characters |
| **Wallet authentication** | Optional. A signature creates or identifies a local profile |
| **Browser pairing** | A six-digit code with local approval, origin binding, hashed storage, expiry, and revocation |
| **Origin policy** | YveChat Engine accepts the exact origins it is configured with. Wildcards are refused |
| **Capability filters** | Models filtered by what they can do, so capability is visible before selection |
***
### Experimental
| Capability | Detail |
| ------------------------------- | --------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- |
| **OpenAI-compatible endpoints** | Configured manually with a base URL. Chat is served conservatively, and capabilities are never guessed from a model name |
| **vLLM** | Reached through the same OpenAI-compatible adapter, suited to machines with GPU infrastructure |
| **Page reading tools** | `open_url`, `find_on_page`, `extract_article`, and `read_pdf` are implemented in the tool gateway and exposed in its catalogue. Chat orchestration currently uses search, so these are reachable through the gateway ahead of full chat integration |
***
### Planned
| Capability | Detail |
| ------------------------------------- | ------------------------------------------------------------------------------------------------------------------------ |
| **Hosted model providers** | Full credential management for the major providers |
| **MCP connections** | Model Context Protocol servers with per-tool permission control |
| **Sandboxed code execution** | Named outside the current MVP trust boundary |
| **Local file and document retrieval** | Attachment and retrieval over local documents |
| **Calculator and date handling** | Deterministic utilities computed by code |
| **Weather, finance, location** | Specialized adapters through explicit providers |
| **External storage** | A user-configurable location for the conversation database |
| **Public installers** | Signed and notarized builds with automatic updates. Unsigned preview builds with published checksums are available today |
***
### The Tool Consent Boundary
Read-only tools show their activity and their results. Any tool that writes, sends, purchases, changes files, or authenticates to an account requires confirmation before it runs.
MCP server installation and arbitrary code execution sit outside the current trust boundary by design. Full detail is in [Tool Gateway](/wp/capabilities/tool-gateway.md).
***
### What every capability depends on
YveChat Engine. The hosted interface is a static browser application, and it discovers no models on its own. Installing and pairing YveChat Engine is the step that makes every row above reachable.
[Getting Started](/wp/getting-started.md) covers that setup.
---
# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.
## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.
Perform an HTTP GET request on the current page URL with the `ask` query parameter, and the optional `goal` query parameter:
```
GET https://yvechat.gitbook.io/wp/capabilities.md?ask=&goal=
```
`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is optional and describes the broader end goal you are ultimately trying to accomplish on behalf of the user. GitBook uses it to tailor the answer towards what is most useful for that goal.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.
Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.