A knowledge instance is a running deployment of app-mediakit-knowledge — the wiki server that renders the platform's documentation and project wikis. Deploying an instance means building the binary, writing a knowledge.toml configuration file pointing at the content repositories, and starting the service. This guide covers deploying a single-instance wiki from a local content path.

For the three-instance model (documentation, projects, corporate), see app-mediakit-knowledge. For declarative content mounts, see use-knowledge-mounts.

Prerequisites

• The app-mediakit-knowledge binary built from the monorepo (see install-toolchain)
• One or more media-knowledge-* content repository clones on the deployment host
• A deployment port (e.g., 9090) that is not occupied by another service
• A terminal session on the host

Step 1: Locate the content repository

The wiki server reads Markdown from a local path. Identify the path to the content repository you want to serve:

ls ~/Foundry/clones/project-editorial/media-knowledge-documentation/

The repository must contain an index.md at root and category subdirectories with _index.md landing pages. If the content has not yet been cloned, clone it first:

git clone git@github.com:pointsav/media-knowledge-documentation.git

Step 2: Write the configuration file

Create knowledge.toml in the deployment directory:

[server]
port = 9090
bind = "0.0.0.0"

[content]
primary_path = "/path/to/media-knowledge-documentation"
instance_name = "documentation"

[search]
enabled = true
index_path = "/var/lib/knowledge/search-index"

[auth]
enabled = false   # set true + configure MBA if editor UI is needed

primary_path must be an absolute path to a media-knowledge-* clone. instance_name appears in the rendered header and in log messages — use a short, descriptive value.

Step 3: Build or locate the binary

If you have not already built the binary, build it from the monorepo:

cd ~/Foundry/clones/<your-archive>/pointsav-monorepo
cargo build -p app-mediakit-knowledge --release

The binary appears at target/release/app-mediakit-knowledge. Copy it to the deployment host if they are different machines.

Step 4: Start the instance

Run the binary with the configuration file:

app-mediakit-knowledge --config knowledge.toml

Or start it as a systemd service using the provided unit file template (if available in the deployment). The service logs startup messages: it reads the content path, builds the search index, and starts listening on the configured port.

Step 5: Verify the instance is serving

Fetch the wiki home page:

curl -s http://127.0.0.1:9090/ | head -20

The response should contain HTML for the wiki home page rendered from index.md. Fetch a category page to confirm article routing:

curl -s http://127.0.0.1:9090/architecture/ | grep '<title>'

If any page returns a 404, check that primary_path in knowledge.toml points to a directory containing the expected category folder.

Key takeaways

• The wiki server reads content directly from the local file path — changes to the content repository appear immediately without a service restart
• instance_name in the configuration is the label used in logs and the rendered header
• The search index is built at startup and lives at index_path — changing this path requires clearing the old index directory
• Three separate instances with three content paths serve documentation, projects, and corporate wikis independently

See also

• app-mediakit-knowledge — the wiki server architecture and the three-instance model
• use-knowledge-mounts — mounting content from multiple repositories into one instance
• install-toolchain — building the binary from the monorepo source
• self-host-a-deployment — the broader deployment procedure of which this is one component
• federate-archives-via-content-mounts — how to serve content from multiple archives in one instance