outo-llms 0.1.0__tar.gz

outo_llms-0.1.0/.gitignore

@@ -0,0 +1,14 @@
+__pycache__/
+*.py[cod]
+*.egg-info/
+.eggs/
+build/
+dist/
+.venv/
+.venv*/
+.env
+.pytest_cache/
+.mypy_cache/
+.ruff_cache/
+.DS_Store
+.omo/

outo_llms-0.1.0/LICENSE

@@ -0,0 +1,201 @@
+                                 Apache License
+                           Version 2.0, January 2004
+                        http://www.apache.org/licenses/
+
+   TERMS AND CONDITIONS FOR USE, REPRODUCTION, AND DISTRIBUTION
+
+   1. Definitions.
+
+      "License" shall mean the terms and conditions for use, reproduction,
+      and distribution as defined by Sections 1 through 9 of this document.
+
+      "Licensor" shall mean the copyright owner or entity authorized by
+      the copyright owner that is granting the License.
+
+      "Legal Entity" shall mean the union of the acting entity and all
+      other entities that control, are controlled by, or are under common
+      control with that entity. For the purposes of this definition,
+      "control" means (i) the power, direct or indirect, to cause the
+      direction or management of such entity, whether by contract or
+      otherwise, or (ii) ownership of fifty percent (50%) or more of the
+      outstanding shares, or (iii) beneficial ownership of such entity.
+
+      "You" (or "Your") shall mean an individual or Legal Entity
+      exercising permissions granted by this License.
+
+      "Source" form shall mean the preferred form for making modifications,
+      including but not limited to software source code, documentation
+      source, and configuration files.
+
+      "Object" form shall mean any form resulting from mechanical
+      transformation or translation of a Source form, including but
+      not limited to compiled object code, generated documentation,
+      and conversions to other media types.
+
+      "Work" shall mean the work of authorship, whether in Source or
+      Object form, made available under the License, as indicated by a
+      copyright notice that is included in or attached to the work
+      (an example is provided in the Appendix below).
+
+      "Derivative Works" shall mean any work, whether in Source or Object
+      form, that is based on (or derived from) the Work and for which the
+      editorial revisions, annotations, elaborations, or other modifications
+      represent, as a whole, an original work of authorship. For the purposes
+      of this License, Derivative Works shall not include works that remain
+      separable from, or merely link (or bind by name) to the interfaces of,
+      the Work and Derivative Works thereof.
+
+      "Contribution" shall mean any work of authorship, including
+      the original version of the Work and any modifications or additions
+      to that Work or Derivative Works thereof, that is intentionally
+      submitted to Licensor for inclusion in the Work by the copyright owner
+      or by an individual or Legal Entity authorized to submit on behalf of
+      the copyright owner. For the purposes of this definition, "submitted"
+      means any form of electronic, verbal, or written communication sent
+      to the Licensor or its representatives, including but not limited to
+      communication on electronic mailing lists, source code control systems,
+      and issue tracking systems that are managed by, or on behalf of, the
+      Licensor for the purpose of discussing and improving the Work, but
+      excluding communication that is conspicuously marked or otherwise
+      designated in writing by the copyright owner as "Not a Contribution."
+
+      "Contributor" shall mean Licensor and any individual or Legal Entity
+      on behalf of whom a Contribution has been received by Licensor and
+      subsequently incorporated within the Work.
+
+   2. Grant of Copyright License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      copyright license to reproduce, prepare Derivative Works of,
+      publicly display, publicly perform, sublicense, and distribute the
+      Work and such Derivative Works in Source or Object form.
+
+   3. Grant of Patent License. Subject to the terms and conditions of
+      this License, each Contributor hereby grants to You a perpetual,
+      worldwide, non-exclusive, no-charge, royalty-free, irrevocable
+      (except as stated in this section) patent license to make, have made,
+      use, offer to sell, sell, import, and otherwise transfer the Work,
+      where such license applies only to those patent claims licensable
+      by such Contributor that are necessarily infringed by their
+      Contribution(s) alone or by combination of their Contribution(s)
+      with the Work to which such Contribution(s) was submitted. If You
+      institute patent litigation against any entity (including a
+      cross-claim or counterclaim in a lawsuit) alleging that the Work
+      or a Contribution incorporated within the Work constitutes direct
+      or contributory patent infringement, then any patent licenses
+      granted to You under this License for that Work shall terminate
+      as of the date such litigation is filed.
+
+   4. Redistribution. You may reproduce and distribute copies of the
+      Work or Derivative Works thereof in any medium, with or without
+      modifications, and in Source or Object form, provided that You
+      meet the following conditions:
+
+      (a) You must give any other recipients of the Work or
+          Derivative Works a copy of this License; and
+
+      (b) You must cause any modified files to carry prominent notices
+          stating that You changed the files; and
+
+      (c) You must retain, in the Source form of any Derivative Works
+          that You distribute, all copyright, patent, trademark, and
+          attribution notices from the Source form of the Work,
+          excluding those notices that do not pertain to any part of
+          the Derivative Works; and
+
+      (d) If the Work includes a "NOTICE" text file as part of its
+          distribution, then any Derivative Works that You distribute must
+          include a readable copy of the attribution notices contained
+          within such NOTICE file, excluding those notices that do not
+          pertain to any part of the Derivative Works, in at least one
+          of the following places: within a NOTICE text file distributed
+          as part of the Derivative Works; within the Source form or
+          documentation, if provided along with the Derivative Works; or,
+          within a display generated by the Derivative Works, if and
+          wherever such third-party notices normally appear. The contents
+          of the NOTICE file are for informational purposes only and
+          do not modify the License. You may add Your own attribution
+          notices within Derivative Works that You distribute, alongside
+          or as an addendum to the NOTICE text from the Work, provided
+          that such additional attribution notices cannot be construed
+          as modifying the License.
+
+      You may add Your own copyright statement to Your modifications and
+      may provide additional or different license terms and conditions
+      for use, reproduction, or distribution of Your modifications, or
+      for any such Derivative Works as a whole, provided Your use,
+      reproduction, and distribution of the Work otherwise complies with
+      the conditions stated in this License.
+
+   5. Submission of Contributions. Unless You explicitly state otherwise,
+      any Contribution intentionally submitted for inclusion in the Work
+      by You to the Licensor shall be under the terms and conditions of
+      this License, without any additional terms or conditions.
+      Notwithstanding the above, nothing herein shall supersede or modify
+      the terms of any separate license agreement you may have executed
+      with Licensor regarding such Contributions.
+
+   6. Trademarks. This License does not grant permission to use the trade
+      names, trademarks, service marks, or product names of the Licensor,
+      except as required for reasonable and customary use in describing the
+      origin of the Work and reproducing the content of the NOTICE file.
+
+   7. Disclaimer of Warranty. Unless required by applicable law or
+      agreed to in writing, Licensor provides the Work (and each
+      Contributor provides its Contributions) on an "AS IS" BASIS,
+      WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or
+      implied, including, without limitation, any warranties or conditions
+      of TITLE, NON-INFRINGEMENT, MERCHANTABILITY, or FITNESS FOR A
+      PARTICULAR PURPOSE. You are solely responsible for determining the
+      appropriateness of using or redistributing the Work and assume any
+      risks associated with Your exercise of permissions under this License.
+
+   8. Limitation of Liability. In no event and under no legal theory,
+      whether in tort (including negligence), contract, or otherwise,
+      unless required by applicable law (such as deliberate and grossly
+      negligent acts) or agreed to in writing, shall any Contributor be
+      liable to You for damages, including any direct, indirect, special,
+      incidental, or consequential damages of any character arising as a
+      result of this License or out of the use or inability to use the
+      Work (including but not limited to damages for loss of goodwill,
+      work stoppage, computer failure or malfunction, or any and all
+      other commercial damages or losses), even if such Contributor
+      has been advised of the possibility of such damages.
+
+   9. Accepting Warranty or Additional Liability. While redistributing
+      the Work or Derivative Works thereof, You may choose to offer,
+      and charge a fee for, acceptance of support, warranty, indemnity,
+      or other liability obligations and/or rights consistent with this
+      License. However, in accepting such obligations, You may act only
+      on Your own behalf and on Your sole responsibility, not on behalf
+      of any other Contributor, and only if You agree to indemnify,
+      defend, and hold each Contributor harmless for any liability
+      incurred by, or claims asserted against, such Contributor by reason
+      of your accepting any such warranty or additional liability.
+
+   END OF TERMS AND CONDITIONS
+
+   APPENDIX: How to apply the Apache License to your work.
+
+      To apply the Apache License to your work, attach the following
+      boilerplate notice, with the fields enclosed by brackets "[]"
+      replaced with your own identifying information. (Don't include
+      the brackets!)  The text should be enclosed in the appropriate
+      comment syntax for the file format. We also recommend that a
+      file or class name and description of purpose be included on the
+      same "printed page" as the copyright notice for easier
+      identification within third-party archives.
+
+   Copyright 2026 outo-llms contributors
+
+   Licensed under the Apache License, Version 2.0 (the "License");
+   you may not use this file except in compliance with the License.
+   You may obtain a copy of the License at
+
+       http://www.apache.org/licenses/LICENSE-2.0
+
+   Unless required by applicable law or agreed to in writing, software
+   distributed under the License is distributed on an "AS IS" BASIS,
+   WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+   See the License for the specific language governing permissions and
+   limitations under the License.

outo_llms-0.1.0/PKG-INFO

@@ -0,0 +1,146 @@
+Metadata-Version: 2.4
+Name: outo-llms
+Version: 0.1.0
+Summary: Deploy local LLMs behind your own managed API server - vLLM or llama.cpp in isolated environments, with API keys, workspaces, and usage tracking.
+Project-URL: Homepage, https://github.com/llaa33219/outo-llms
+Project-URL: Repository, https://github.com/llaa33219/outo-llms
+Project-URL: Issues, https://github.com/llaa33219/outo-llms/issues
+Author: outo-llms contributors
+License: Apache-2.0
+License-File: LICENSE
+Keywords: api-server,inference,llama.cpp,llm,self-hosted,vllm
+Classifier: Development Status :: 3 - Alpha
+Classifier: Environment :: Console
+Classifier: Intended Audience :: Developers
+Classifier: License :: OSI Approved :: Apache Software License
+Classifier: Operating System :: OS Independent
+Classifier: Programming Language :: Python :: 3
+Classifier: Programming Language :: Python :: 3 :: Only
+Classifier: Topic :: Scientific/Engineering :: Artificial Intelligence
+Requires-Python: >=3.10
+Requires-Dist: cryptography>=42
+Requires-Dist: fastapi>=0.110
+Requires-Dist: httpx>=0.27
+Requires-Dist: platformdirs>=4
+Requires-Dist: pydantic>=2
+Requires-Dist: rich>=13
+Requires-Dist: typer>=0.12
+Requires-Dist: uvicorn[standard]>=0.29
+Description-Content-Type: text/markdown
+
+# outo-llms
+
+Deploy local LLMs behind your own managed, OpenAI-compatible API server.
+
+vLLM or llama.cpp runs in an **isolated environment** managed by outo-llms,
+kept apart from the Python environments you already have. Users sign up,
+receive API keys, organize work into workspaces, and every request is
+metered per workspace. Open the dashboard, point your OpenAI SDK at the
+URL, and the rest is just an HTTP call.
+
+Full documentation lives in [`docs/`](docs/index.md).
+
+## Principles
+
+1. **Fragmentation** - small, single-purpose modules.
+2. **Fluidity** - features can be added or removed as needs change.
+3. **Simplicity** - a handful of commands and a small API cover the common workflow.
+4. **Automation** - one guided command gets you a working deployment.
+5. **Explicitness** - every automated action is announced and logged; the
+   system is never touched without your knowledge.
+
+## Features
+
+- **Isolated engine virtualenvs.** vLLM and llama.cpp install into their
+  own virtual environments, never into the system interpreter.
+- **Choice of engines.** Bring Hugging Face models through vLLM, or GGUF
+  models through llama.cpp, and switch between them with one command.
+- **Account, keys, and workspaces.** Open signup issues an `outo_sk_` API
+  key, creates a `default` workspace, and lets each user create more.
+- **Per-workspace usage metering.** Token usage is recorded by workspace
+  and surfaced at `GET /v1/usage`.
+- **OpenAI-compatible proxy.** `POST /v1/chat/completions`,
+  `POST /v1/completions`, and `GET /v1/models` speak the same shapes
+  clients already know.
+- **Optional HTTPS.** A self-signed certificate is generated under
+  `data/certs/` when you ask for it.
+- **Explicit automation with an action log.** Every setup step is
+  announced, confirmed when destructive, and appended to `actions.log`.
+- **Dashboard and Swagger UI.** Visit the root URL for a tiny status
+  page, or `/docs` for the full OpenAPI explorer.
+
+## Install
+
+```bash
+pip install outo-llms
+```
+
+Prefer an isolated install? `pipx install outo-llms` works too.
+
+## Quickstart
+
+```bash
+outo-llms setup                              # interactive, fully explicit setup
+curl -s -X POST http://127.0.0.1:8611/v1/account/signup \
+  -H 'Content-Type: application/json' \
+  -d '{"username": "me"}'                    # returns an api_key + default workspace
+outo-llms models add tinyllama               # register a Hugging Face model
+curl -s http://127.0.0.1:8611/v1/chat/completions \
+  -H "Authorization: Bearer outo_sk_..." \
+  -H 'Content-Type: application/json' \
+  -d '{"model": "tinyllama", "messages": [{"role": "user", "content": "hi"}]}'
+curl -s http://127.0.0.1:8611/v1/usage \
+  -H "Authorization: Bearer outo_sk_..."    # per-workspace token accounting
+```
+
+The setup wizard prints the exact base URL it configured (HTTP or HTTPS),
+along with the path to the action log, and writes a short summary you can
+copy into your shell history.
+
+## Commands
+
+| Command | Purpose |
+| --- | --- |
+| `outo-llms setup` | Automated, explicit server setup (engine, HTTPS, firewall, launch) |
+| `outo-llms models add <name>` | Register a model in the registry |
+| `outo-llms models list` | Show every registered model |
+| `outo-llms models remove <name>` | Unregister a model (asks first) |
+| `outo-llms engine list` | List known engines and their installed state |
+| `outo-llms engine use <name>` | Switch the active engine (llamacpp or vllm) |
+| `outo-llms engine install [name]` | Install an engine into its isolated virtualenv |
+| `outo-llms engine status` | Show the active engine's runtime status |
+| `outo-llms start` | Start the API server in the background |
+| `outo-llms stop` | Stop the background API server |
+| `outo-llms restart` | Restart the API server |
+| `outo-llms status` | Show server, engine, and path status |
+| `outo-llms reset` | Wipe everything back to factory state (asks twice) |
+| `outo-llms version` | Print the version |
+
+## How it works
+
+The outo-llms server (`python -m outo_llms.server`) binds `127.0.0.1:8611`
+by default and exposes an OpenAI-compatible HTTP API. Engines are
+internal services that bind their own loopback ports (llama.cpp on 8612,
+vLLM on 8613); clients never reach them directly.
+
+When a request hits `POST /v1/chat/completions` or `POST /v1/completions`,
+the server looks up the requested model in the registry, asks the engine
+manager to ensure the right engine is running with that model, then
+forwards the request. The active engine streams or returns its response
+unchanged, and the server records prompt and completion tokens against
+the calling workspace.
+
+Every state change, install step, and external command goes through
+`outo_llms.core.consent`, which announces what is about to happen, asks
+for confirmation on destructive actions, and writes a timestamped line to
+`logs/actions.log`. The dashboard at the root URL and Swagger UI at
+`/docs` round out the picture.
+
+## Documentation
+
+- [`docs/index.md`](docs/index.md) — entry point for installation, API,
+  configuration, operations, and testing guides.
+
+## License
+
+Apache License 2.0. See [`LICENSE`](LICENSE).

outo_llms-0.1.0/README.md

@@ -0,0 +1,116 @@
+# outo-llms
+
+Deploy local LLMs behind your own managed, OpenAI-compatible API server.
+
+vLLM or llama.cpp runs in an **isolated environment** managed by outo-llms,
+kept apart from the Python environments you already have. Users sign up,
+receive API keys, organize work into workspaces, and every request is
+metered per workspace. Open the dashboard, point your OpenAI SDK at the
+URL, and the rest is just an HTTP call.
+
+Full documentation lives in [`docs/`](docs/index.md).
+
+## Principles
+
+1. **Fragmentation** - small, single-purpose modules.
+2. **Fluidity** - features can be added or removed as needs change.
+3. **Simplicity** - a handful of commands and a small API cover the common workflow.
+4. **Automation** - one guided command gets you a working deployment.
+5. **Explicitness** - every automated action is announced and logged; the
+   system is never touched without your knowledge.
+
+## Features
+
+- **Isolated engine virtualenvs.** vLLM and llama.cpp install into their
+  own virtual environments, never into the system interpreter.
+- **Choice of engines.** Bring Hugging Face models through vLLM, or GGUF
+  models through llama.cpp, and switch between them with one command.
+- **Account, keys, and workspaces.** Open signup issues an `outo_sk_` API
+  key, creates a `default` workspace, and lets each user create more.
+- **Per-workspace usage metering.** Token usage is recorded by workspace
+  and surfaced at `GET /v1/usage`.
+- **OpenAI-compatible proxy.** `POST /v1/chat/completions`,
+  `POST /v1/completions`, and `GET /v1/models` speak the same shapes
+  clients already know.
+- **Optional HTTPS.** A self-signed certificate is generated under
+  `data/certs/` when you ask for it.
+- **Explicit automation with an action log.** Every setup step is
+  announced, confirmed when destructive, and appended to `actions.log`.
+- **Dashboard and Swagger UI.** Visit the root URL for a tiny status
+  page, or `/docs` for the full OpenAPI explorer.
+
+## Install
+
+```bash
+pip install outo-llms
+```
+
+Prefer an isolated install? `pipx install outo-llms` works too.
+
+## Quickstart
+
+```bash
+outo-llms setup                              # interactive, fully explicit setup
+curl -s -X POST http://127.0.0.1:8611/v1/account/signup \
+  -H 'Content-Type: application/json' \
+  -d '{"username": "me"}'                    # returns an api_key + default workspace
+outo-llms models add tinyllama               # register a Hugging Face model
+curl -s http://127.0.0.1:8611/v1/chat/completions \
+  -H "Authorization: Bearer outo_sk_..." \
+  -H 'Content-Type: application/json' \
+  -d '{"model": "tinyllama", "messages": [{"role": "user", "content": "hi"}]}'
+curl -s http://127.0.0.1:8611/v1/usage \
+  -H "Authorization: Bearer outo_sk_..."    # per-workspace token accounting
+```
+
+The setup wizard prints the exact base URL it configured (HTTP or HTTPS),
+along with the path to the action log, and writes a short summary you can
+copy into your shell history.
+
+## Commands
+
+| Command | Purpose |
+| --- | --- |
+| `outo-llms setup` | Automated, explicit server setup (engine, HTTPS, firewall, launch) |
+| `outo-llms models add <name>` | Register a model in the registry |
+| `outo-llms models list` | Show every registered model |
+| `outo-llms models remove <name>` | Unregister a model (asks first) |
+| `outo-llms engine list` | List known engines and their installed state |
+| `outo-llms engine use <name>` | Switch the active engine (llamacpp or vllm) |
+| `outo-llms engine install [name]` | Install an engine into its isolated virtualenv |
+| `outo-llms engine status` | Show the active engine's runtime status |
+| `outo-llms start` | Start the API server in the background |
+| `outo-llms stop` | Stop the background API server |
+| `outo-llms restart` | Restart the API server |
+| `outo-llms status` | Show server, engine, and path status |
+| `outo-llms reset` | Wipe everything back to factory state (asks twice) |
+| `outo-llms version` | Print the version |
+
+## How it works
+
+The outo-llms server (`python -m outo_llms.server`) binds `127.0.0.1:8611`
+by default and exposes an OpenAI-compatible HTTP API. Engines are
+internal services that bind their own loopback ports (llama.cpp on 8612,
+vLLM on 8613); clients never reach them directly.
+
+When a request hits `POST /v1/chat/completions` or `POST /v1/completions`,
+the server looks up the requested model in the registry, asks the engine
+manager to ensure the right engine is running with that model, then
+forwards the request. The active engine streams or returns its response
+unchanged, and the server records prompt and completion tokens against
+the calling workspace.
+
+Every state change, install step, and external command goes through
+`outo_llms.core.consent`, which announces what is about to happen, asks
+for confirmation on destructive actions, and writes a timestamped line to
+`logs/actions.log`. The dashboard at the root URL and Swagger UI at
+`/docs` round out the picture.
+
+## Documentation
+
+- [`docs/index.md`](docs/index.md) — entry point for installation, API,
+  configuration, operations, and testing guides.
+
+## License
+
+Apache License 2.0. See [`LICENSE`](LICENSE).

outo_llms-0.1.0/docs/architecture.md

@@ -0,0 +1,151 @@
+# Architecture
+
+This page describes the module boundaries, the request flow, and the extension points. The implementation is deliberately small. Each concern lives in one module.
+
+## Module map
+
+```text
+src/outo_llms/
+├── __init__.py             package metadata and version
+├── __main__.py             `python -m outo_llms` placeholder
+├── core/
+│   ├── paths.py            XDG-style paths and directory creation
+│   ├── config.py           JSON-backed server and engine configuration
+│   ├── consent.py          announce, confirm, confirm_twice, log_action, run_system
+│   ├── process.py          server start, stop, restart, pid helpers, port polling
+│   └── certs.py            self-signed certificate generation for HTTPS
+├── engines/
+│   ├── base.py             EngineAdapter contract, ModelRef, adapter registry
+│   ├── llamacpp.py         llama.cpp adapter and serve argv
+│   ├── vllm.py             vLLM adapter and serve argv
+│   └── manager.py          EngineManager: install, use, ensure_running, stop
+├── server/
+│   ├── __main__.py         uvicorn entry point: `python -m outo_llms.server`
+│   ├── app.py              FastAPI factory, lifespan, /healthz, /
+│   ├── db.py               SQLite connection, schema, init_db, utcnow
+│   ├── accounts.py         users, workspaces, key issue/revoke/verify
+│   ├── registry.py         model registry CRUD
+│   ├── usage.py            usage record and aggregate summary
+│   ├── schemas.py          Pydantic v2 request and response models
+│   ├── deps.py             OpenAI-style error helper, WorkspaceDep auth dependency
+│   ├── proxy.py            OpenAI-compatible proxy with usage accounting
+│   └── routes/
+│       ├── account.py      POST /v1/account/signup, GET /v1/account/me
+│       ├── workspaces.py   POST /v1/workspaces, GET /v1/workspaces
+│       ├── keys.py         POST/GET /v1/workspaces/{name}/keys, DELETE /v1/keys/{key_id}
+│       └── usage.py        GET /v1/usage
+└── cli/
+    ├── app.py              Typer application, command registration
+    └── commands/
+        ├── setup.py        outo-llms setup
+        ├── start.py        outo-llms start
+        ├── stop.py         outo-llms stop
+        ├── restart.py      outo-llms restart
+        ├── status.py       outo-llms status
+        ├── reset.py        outo-llms reset
+        ├── version.py      outo-llms version
+        ├── models.py       outo-llms models add/list/remove
+        └── engine.py       outo-llms engine list/use/install/status
+```
+
+## Layered responsibilities
+
+The implementation is layered:
+
+* `core/` is shared infrastructure. It has no knowledge of engine adapters, accounts, or HTTP routes.
+* `engines/` knows about adapters and how to run them. It calls `core/` for paths, consent, configuration, and process control.
+* `server/` knows about HTTP, persistence, accounts, the proxy, and the OpenAI-compatible surface. It calls `engines/` lazily for inference and `core/` for shared helpers.
+* `cli/` knows about the user command surface. It calls into the same modules as the server.
+
+The boundaries mean a change to a CLI command does not change the server, and a change to an engine adapter does not change accounts or the CLI.
+
+## Request flow
+
+```text
+client                outo-llms API server                EngineManager       engine subprocess
+  │                          │                                  │                    │
+  │ HTTP /v1/chat/...        │                                  │                    │
+  │ (Bearer outo_sk_...)     │                                  │                    │
+  │ ────────────────────────▶                                  │                    │
+  │                          │ require_workspace():            │                    │
+  │                          │   parse header, verify key      │                    │
+  │                          │   → WorkspaceContext            │                    │
+  │                          │ resolve model in registry       │                    │
+  │                          │ ensure_running(model_ref) ─────▶│                    │
+  │                          │                                  │ start / reuse      │
+  │                          │                                  │ engine process ───▶│
+  │                          │                                  │ poll /v1/models    │
+  │                          │                                  │ until 200          │
+  │                          │ ensure_running returns base_url │                    │
+  │                          │ proxy.forward → upstream URL    │                    │
+  │                          │ ────────────────────────────────▶ ──────────────────▶│
+  │                          │                                  │                    │ run inference
+  │                          │ ◀──────────────────────────────── ───────────────────│
+  │                          │ record usage (non-stream or     │                    │
+  │                          │   final SSE event with usage)   │                    │
+  │ ◀────────────────────────                                  │                    │
+  │ OpenAI-style response or SSE stream                        │                    │
+```
+
+For `chat/completions` and `completions`, the proxy does the following:
+
+1. Parse and validate the JSON body.
+2. Resolve the `model` field against the registry. Missing `model` is HTTP `400`. Unknown model is HTTP `404` with `code: "model_not_found"`.
+3. Call `EngineManager.ensure_running(model_ref)` to start or reuse the engine. A missing engine installation or an unsupported kind is HTTP `502`.
+4. Forward the request to the upstream OpenAI-compatible URL on `127.0.0.1:<engine-port>`.
+5. For non-streaming responses, parse JSON, read `usage.prompt_tokens` and `usage.completion_tokens`, and record a usage row.
+6. For streaming responses, set `stream_options.include_usage: true`, stream SSE bytes, then inspect the assembled text for a usage-bearing event and record usage best effort.
+
+The same dependency authenticates every authenticated route. Errors return through the OpenAI-style error helper in `server/deps.py`.
+
+## Process model
+
+```text
+shell
+└── outo-llms start    (this process exits after subprocess.Popen + PID write)
+    └── python -m outo_llms.server        (API server, detached via start_new_session=True)
+        ├── uvicorn loop
+        │   ├── /healthz, /, /docs
+        │   ├── /v1/account/signup, /v1/account/me
+        │   ├── /v1/workspaces, /v1/workspaces/{name}/keys
+        │   ├── /v1/keys/{key_id}
+        │   ├── /v1/usage
+        │   └── /v1/models, /v1/chat/completions, /v1/completions
+        └── EngineManager.ensure_running on demand
+            └── python -m llama_cpp.server (or vllm.entrypoints.openai.api_server)
+```
+
+The API server is a detached process with output redirected to `data/logs/server.log`. Its PID is stored in `data/server.pid`. `outo-llms stop` reads the PID file, sends SIGTERM, waits for the process to exit, and sends SIGKILL if needed.
+
+Engine processes follow the same pattern with PID, port, and model name written directly under `data/engines/`. `EngineManager` reuses a running engine when the same model is requested, and it stops and restarts the engine when a different model is requested or when a stale PID is detected.
+
+The `__main__.py` for `outo_llms.server` calls `uvicorn.run(...)` with the configured host and port. If HTTPS is enabled, it asks `core/certs.py` for a self-signed pair and passes them to Uvicorn.
+
+## How fragmentation and fluidity show up
+
+* **Adding an engine.** Create a new module under `src/outo_llms/engines/` with a class that extends `EngineAdapter`. Add its name to `adapter_names()` and the lookup in `get_adapter()`. The manager, server, and CLI pick it up automatically.
+* **Adding a CLI command.** Create a new module under `src/outo_llms/cli/commands/` with a Typer function. Register it in `src/outo_llms/cli/app.py`. For a command group, create a Typer app and use `app.add_typer`.
+* **Adding an HTTP route.** Create a new module under `src/outo_llms/server/routes/` with a router. Include it in `src/outo_llms/server/app.py`. Use the `WorkspaceDep` dependency for authenticated routes and the `openai_error` helper for OpenAI-style error responses.
+* **Adding configuration.** Extend a dataclass in `src/outo_llms/core/config.py`. The loader falls back to defaults for missing keys, so older `config.json` files still work.
+
+The [development guide](development.md) walks through each extension in detail.
+
+## Data and persistence
+
+SQLite is the only database. The schema covers five tables: `users`, `workspaces`, `api_keys`, `models`, and `usage`. The schema is created by `db.init_db()` from a single `CREATE TABLE IF NOT EXISTS` script. All timestamps are UTC ISO-8601 strings. Foreign keys are enabled per connection. Queries use parameter binding.
+
+The `EngineManager` keeps no in-memory state between calls. All state lives on disk: `config.json`, engine marker files, engine PID, port, model files, the database, the action log, and engine and server logs.
+
+## What does not change
+
+These are stable contracts documented for users and contributors:
+
+* The `EngineManager` API in `engines/manager.py`.
+* The registry functions `add_model`, `get_model`, `list_models`, and `remove_model` in `server/registry.py`.
+* `db.init_db` in `server/db.py`.
+* The process helpers `start_server`, `stop_server`, and `restart_server` in `core/process.py`.
+* The console entry point `outo_llms.cli.app:main`, exposed as `outo-llms`.
+* The module entry point `python -m outo_llms.server`.
+* The OpenAI-style error shape.
+
+Changes to those contracts require updates to every caller and to the documentation.