diff --git a/website/src/routes/docs/+layout.svelte b/website/src/routes/docs/+layout.svelte index 23aa979..9647819 100644 --- a/website/src/routes/docs/+layout.svelte +++ b/website/src/routes/docs/+layout.svelte @@ -8,9 +8,8 @@ { href: '/docs/installation', label: 'Installation' }, { href: '/docs/usage', label: 'Usage' }, { href: '/docs/configuration', label: 'Configuration' }, - { href: '/docs/providers', label: 'Providers' }, - { href: '/docs/self-hosted', label: 'Self-Hosted' }, - { href: '/docs/examples', label: 'Examples' }, + { href: '/docs/providers', label: 'Providers' }, + { href: '/docs/examples', label: 'Examples' }, { href: '/docs/cli-reference', label: 'CLI Reference' }, ]; diff --git a/website/src/routes/docs/getting-started/+page.svelte b/website/src/routes/docs/getting-started/+page.svelte index 2c4c1d1..7195d0b 100644 --- a/website/src/routes/docs/getting-started/+page.svelte +++ b/website/src/routes/docs/getting-started/+page.svelte @@ -149,6 +149,6 @@ Usage — review modes, output formats, and configuration Configuration — full .cora.yaml reference Providers — setting up OpenAI, Anthropic, Groq, Ollama, and Z.AI - Self-Hosted — running cora with local LLMs and air-gapped environments + CLI Reference — full command documentation diff --git a/website/src/routes/docs/self-hosted/+page.svelte b/website/src/routes/docs/self-hosted/+page.svelte deleted file mode 100644 index da2e109..0000000 --- a/website/src/routes/docs/self-hosted/+page.svelte +++ /dev/null @@ -1,217 +0,0 @@ - - Self-Hosted — cora docs - - - - -

- Self-Hosted -

- -
-

Why Self-Hosted?

-

cora is designed to run entirely on your infrastructure. Self-hosting gives you:

-
-
- • - Privacy — Your code never leaves your network -
-
- • - Compliance — Meet regulatory requirements for code review in sensitive environments -
-
- • - Air-gapped — Works in fully offline environments with local LLMs -
-
- • - No vendor lock-in — Switch providers or models at any time -
-
-
- -
-

Using Ollama (Local LLM)

-

Ollama lets you run LLMs locally. cora works with any Ollama model that supports chat completions.

- -
-
1
-
- Install Ollama — Follow the instructions at ollama.ai -
-
- -
-
2
-
- Pull a model — For code review, use ollama pull codellama -
-
- -
-
3
-
- Configure cora — Point cora to your local Ollama instance -
-
- -
-
- - - -
-
-
# Start Ollama
-
$ ollama serve
-
-
# Pull a code-focused model
-
$ ollama pull codellama
-
-
- -

Configure cora to use Ollama via environment variables:

- -
-
- - - -
-
-
$ CORA_PROVIDER=ollama CORA_BASE_URL=http://localhost:11434/v1 CORA_MODEL=codellama cora review --staged
-
-
- -

Or persist the configuration in .cora.yaml:

- -
-
- - - -
-
-
# .cora.yaml
-
provider: ollama
-
base_url: http://localhost:11434/v1
-
model: codellama
-
-
- -

To authenticate with Ollama (no API key needed for local use):

- -
-
- - - -
-
-
$ cora auth login --provider ollama --base-url http://localhost:11434/v1
-
-
-
- -
-

Using with Gitea

-

cora works seamlessly with Gitea repositories. Since cora reads your local git state directly, no Gitea integration or plugin is needed:

- -
-
- • - cora reads git diff output — it works with any git remote (GitHub, Gitea, GitLab, Bitbucket) -
-
- • - No API tokens, webhooks, or server-side configuration required -
-
- • - Works with Gitea's CI/CD pipelines — just add cora review --branch main as a pipeline step -
-
- -
-
- - - -
-
-
# Clone your Gitea repo and review
-
$ git clone gitea.example.com/org/repo.git
-
$ cd repo
-
$ cora init
-
$ cora review --staged
-
-
-
- -
-

Network Isolation

-

For fully offline or air-gapped environments:

- -
-
- • - Download the cora binary and transfer it to the isolated machine -
-
- • - Run Ollama on the same machine with a locally-pulled model -
-
- • - Configure cora to use http://localhost:11434/v1 as the base URL -
-
- • - No outbound network connections are made — everything stays on the local machine -
-
-
- -
-

No Telemetry

-

cora is designed with privacy as a core principle:

- -
-
- • - Zero data collection — cora does not send any data to any server -
-
- • - No analytics — No usage tracking, no metrics, no phone home -
-
- • - Local execution — Everything runs on your machine -
-
- • - The only network traffic — API calls to your configured LLM provider (which you control) -
-
- -

You can verify this by inspecting the source code at github.com/ajianaz/cora-cli. The codebase is open source under the MIT license.

-