Click on "Install Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type @ followed by the MCP server name and your instructions, e.g., "@mcp-kettlelogic search for articles on AI governance"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
A Model Context Protocol (MCP) server that exposes a Kettle Logic site's published content — whitepapers/playbooks ("insights") and industry pages — to any MCP-capable agent (Claude Desktop, IDEs, custom orchestrators).
It is a pure, read-only client of the public website: it fetches content live over HTTP and parses it with the Python standard library. No database, no API keys, no dependency on any backend cluster or LLM. The target site is configurable, so you can point it at your own deployment — or any site with the same shape — and see what it discovers.
Built on the official mcp Python SDK (FastMCP),
speaking the standard stdio transport (and an optional streamable-http
transport for container/Kubernetes deployment).
Engineering note: this repository doubles as a reference for how we build — hexagonal layering, fully-typed OO, dependency injection, and a suite of ratchets (see ARCHITECTURE.md) that enforce the rules in CI.
pytestruns unit tests, real-stdio e2e tests, and the ratchets together.
Tool | Description |
| Search insight articles by title / slug / description. |
| Plain-text overview extracted from an industry page. |
| JSON catalog of every insight article (title, slug, description). |
| JSON list of industries with guidance (name + slug). |
| A single insight article rendered as readable text. |
URI | Description |
| JSON catalog of every insight article. |
| JSON list of industry pages discovered on the site. |
| A single article rendered as readable text. |
Discovery is live: articles from the site's /insights/ index, industries from its
/llms.txt (falling back to crawling /industries/).
Related MCP server: docs-mcp-server
pip install . # Python >= 3.11; pulls in mcp + httpx
mcp-kettlelogic # console script (stdio transport by default)Normally an MCP client launches it. To explore interactively:
npx @modelcontextprotocol/inspector mcp-kettlelogic{
"mcpServers": {
"kettlelogic": {
"command": "mcp-kettlelogic",
"env": { "KETTLELOGIC_BASE_URL": "https://kettlelogic.com" }
}
}
}A hardened instance runs live over streamable-HTTP — point any MCP client at it:
{
"mcpServers": {
"kettlelogic": { "url": "https://kettlelogic.com/mcp" }
}
}This server is listed in the public MCP registry
as com.kettlelogic/mcp-kettlelogic (brand: kettlelogic, not a personal GitHub
handle). server.json is the manifest — it advertises the hosted
streamable-HTTP remote and validates against the 2025-12-11 schema.
The listing is claimed by HTTP domain ownership, not a GitHub login: the registry
fetches https://kettlelogic.com/.well-known/mcp-registry-auth (a public-key challenge
served by the marketing site — web/public/.well-known/mcp-registry-auth in the
kettlelogic repo) and verifies a signature made with our Ed25519 private key.
.github/workflows/publish-registry.yml
publishes automatically on every GitHub Release (and via manual Run workflow).
It runs on a self-hosted runner in our cluster (runs-on: [self-hosted, linux, cluster]) — chosen because GitHub-hosted minutes run out fast, and this way the
publish costs zero quota and won't fail when the month is drained. That runner also
mounts the signing key from an in-cluster Secret (/opt/mcp/key-hex), so the key
never lives in GitHub Actions secrets.
So the normal flow is: bump version.py + server.json version → cut a GitHub
Release → the registry updates itself.
Runner setup is one-time and lives in the private home repo:
deployments/github-runner/ (README there). If the runner is down, re-run from
the Actions tab, or use the local fallback below — it always works with no infra.
If you ever need to publish by hand:
# 1. Get the CLI (prebuilt binary from the registry releases)
curl -L https://github.com/modelcontextprotocol/registry/releases/latest/download/mcp-publisher_linux_amd64.tar.gz | tar xz mcp-publisher
# 2. From the repo root (where server.json lives):
./mcp-publisher validate
./mcp-publisher login http --domain kettlelogic.com \
--private-key "$(openssl pkey -in .secrets/mcp-registry-key.pem -outform DER | tail -c 32 | xxd -p -c 64)"
./mcp-publisher publishThe Ed25519 key authorizes publishing under the com.kettlelogic namespace. It
exists in three places (the private key is intentionally NOT in git):
Local — .secrets/mcp-registry-key.pem (gitignored). Convenience for manual publishes.
Cluster — Secret mcp-registry-key (key key-hex) in the github-runner
namespace. This is what the auto-publish runner uses, and the canonical backup.
Public half — served at https://kettlelogic.com/.well-known/mcp-registry-auth
(committed in the kettlelogic repo as web/public/.well-known/mcp-registry-auth).
Lost the local .pem? Recover the signing key (as hex) straight from the cluster —
publishing only needs the hex, not the pem file:
KEY_HEX="$(kubectl -n github-runner get secret mcp-registry-key -o jsonpath='{.data.key-hex}' | base64 -d)"
./mcp-publisher login http --domain kettlelogic.com --private-key "$KEY_HEX"
./mcp-publisher publishRotate (if the key is ever exposed):
openssl genpkey -algorithm Ed25519 -out .secrets/mcp-registry-key.pem # new key
PUB="$(openssl pkey -in .secrets/mcp-registry-key.pem -pubout -outform DER | tail -c 32 | base64)"
# 1. update the served challenge (in the kettlelogic repo) + redeploy the site:
echo "v=MCPv1; k=ed25519; p=$PUB" > ../kettlelogic/web/public/.well-known/mcp-registry-auth
# 2. update the cluster Secret the runner uses:
kubectl -n github-runner delete secret mcp-registry-key
kubectl -n github-runner create secret generic mcp-registry-key \
--from-literal=key-hex="$(openssl pkey -in .secrets/mcp-registry-key.pem -outform DER | tail -c 32 | xxd -p -c 64)"
# 3. re-publish (new key must be live at the well-known URL first).Versions are immutable in the registry — you can't re-publish an existing version;
bump version.py + server.json first.
Env var | Default | Purpose |
|
| Target site to read. Point it at your own. |
|
|
|
|
| Bind for the http transport. |
| (unset) | If set, serve Prometheus metrics at |
|
| Log level. Logs go to stderr (stdout is the MCP channel). |
| see | Tuning. |
For network deployment the server runs the streamable-http transport.
docker compose -f deploy/docker/docker-compose.yaml up --buildKubernetes manifests live in deploy/k8s/ (Deployment with 2
replicas, readiness/liveness probes, resource bounds, non-root + read-only root FS,
ClusterIP Service, ConfigMap). Apply with:
kubectl apply -k deploy/k8sLogging — structured logs to stderr: operation start/finish + durations,
fetch results, cache hits/misses, errors. Set KETTLELOGIC_LOG_LEVEL.
Metrics — Prometheus /metrics (set KETTLELOGIC_METRICS_PORT):
mcp_operations_total, mcp_errors_total, mcp_http_fetches_total,
mcp_http_errors_total, mcp_cache_total{result}, mcp_op_duration_seconds.
pip install -e ".[dev]"
ruff check src tests # lint + import order
mypy # type check (disallow-untyped-defs)
pytest # unit + e2e + ratchets, coverage gate (fail-under 90%)Coverage runs ~99%. See ARCHITECTURE.md for layering, the request flow, and the full list of ratchets.
MIT — see LICENSE.
Fetch one insight article as readable text (slug, e.g. "control-tower-operations").
| Name | Required | Description | Default |
|---|---|---|---|
| slug | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided, so description carries full burden. It states 'fetch' implying read-only, but does not disclose permissions, error handling, rate limits, or whether the operation is idempotent. Minimal behavioral context.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
One sentence (12 words) with key information front-loaded. No fluff or redundancy; every word adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Adequate for a simple single-parameter tool with an output schema. Covers basic purpose and parameter but lacks usage guidelines and behavioral transparency, leaving gaps for an agent.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 0%, so description must compensate. Provides an example for 'slug' parameter, adding context beyond the schema's type and title, but no further details on slug format or how to obtain it.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Clearly states verb 'Fetch', resource 'insight article', and output format 'readable text'. Distinguishes from siblings like search_articles and list_articles by specifying retrieval by slug with an example.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Implies usage when a specific slug is known (e.g., 'control-tower-operations'), but does not explicitly state when to use this over siblings, nor provides exclusions or prerequisites.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Return a plain-text overview for an industry page (slug, e.g. "retail").
| Name | Required | Description | Default |
|---|---|---|---|
| industry | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations are provided, so the description carries the full burden. It states the return type (plain-text overview) but does not disclose whether the operation is read-only, any authorization needs, rate limits, or side effects. The gap is significant for a tool with zero annotation coverage.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single 11-word sentence that efficiently communicates the tool's purpose and parameter style. No wasted words; every part adds value.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
For a simple retrieval tool with one parameter and an output schema present, the description is mostly adequate. It lacks mention that the slug should be valid (e.g., from list_industries) and error conditions, but the core functionality is covered.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
The single parameter 'industry' has no description in the schema (0% coverage). The description specifies it is a 'slug' and gives an example ('retail'), adding critical meaning beyond the schema's title alone.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description uses a specific verb 'Return' and resource 'overview for an industry page' with a concrete example slug. It clearly distinguishes from siblings like get_article and search_articles by targeting industry overviews.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
The description implies usage for obtaining an industry page overview but provides no explicit when-to-use, when-not-to-use, or references to alternative tools like list_industries. It lacks guidance on context.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List every Kettle Logic insight article (title, slug, description) as JSON.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
The description indicates a simple read operation with no side effects. For a list tool with no parameters and no annotations, it sufficiently explains the behavior. However, it omits potential details like pagination or rate limits, which may be inferred but are not explicit.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
A single sentence with no wasted words, front-loaded with the core action and resource. It is efficient and easy to parse.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given the tool has no parameters and an output schema exists, the description covers the essential purpose and output. It is complete for typical use, though lacking explicit mention of pagination or ordering. Still, it is adequate for a simple list operation.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
There are no parameters, so the description adds value by specifying the return fields (title, slug, description) beyond the empty schema. This exceeds the baseline expectation.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the action (list), resource (every Kettle Logic insight article), and output format (JSON), including specific fields returned (title, slug, description). It effectively distinguishes from sibling tools like get_article or search_articles by implying it returns all articles.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
Usage is implied: use to get all articles. However, there is no explicit guidance on when to use this vs. sibling tools (e.g., search_articles for filtered results) or when not to use it.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
List the industries Kettle Logic publishes guidance for (name + slug) as JSON.
| Name | Required | Description | Default |
|---|---|---|---|
No parameters | |||
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
Without annotations, the description carries the burden. It declares the tool as read-only (list) and specifies JSON output, which is transparent. No contradictions or missing behavioral implications.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
One sentence with 16 words, no redundancy, front-loaded with the action 'List'. Every word earns its place.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given no parameters and an existing output schema, the description is sufficient. It covers what the tool does and the return format.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
No parameters exist, so schema coverage is 100%. The description does not need to add param info. Baseline 4 applies.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
The description clearly states the verb 'List', the resource 'industries', and the output format 'name + slug as JSON'. It is distinct from sibling tools like get_article or search_articles.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit guidance on when to use this tool versus alternatives, though the purpose is clear. It implies usage for listing all industries, but does not mention when-not or point to siblings like get_industry_overview.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Search Kettle Logic insight articles by title, slug or description.
| Name | Required | Description | Default |
|---|---|---|---|
| limit | No | ||
| query | Yes |
| Name | Required | Description |
|---|---|---|
| result | Yes |
Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?
No annotations provided; description only mentions search fields. Does not disclose case sensitivity, partial matching, pagination behavior, or any further constraints, leaving agent uncertain about search mechanics.
Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.
Is the description appropriately sized, front-loaded, and free of redundancy?
Single sentence, no filler, all words carry meaning. Efficient and well-structured.
Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.
Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?
Given output schema exists, description does not need to explain return values. Covers search scope and parameter hint, but lacks details on result ordering, error handling, or pagination. Adequate but not exhaustive.
Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.
Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?
Schema coverage is 0%, so description must compensate. It clarifies that 'query' searches title, slug, and description, but does not explain the 'limit' parameter beyond its schema default. Partial addition but incomplete.
Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.
Does the description clearly state what the tool does and how it differs from similar tools?
Description uses specific verb 'Search', identifies resource 'articles', and lists fields 'title, slug or description', clearly distinguishing from sibling tools like get_article or list_articles.
Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.
Does the description explain when to use this tool, when not to, or what alternatives exist?
No explicit when-to-use or alternatives mentioned. Implication is clear for searching but lacks guidance on when to use sibling tools like get_article for exact article retrieval or list_articles for unfiltered listing.
Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.
Recent tool additions, removals, and schema changes observed during successful MCP inspections. Dates show when Glama detected each change.
v0.1.0get_articleget_industry_overviewlist_articleslist_industriessearch_articlesEach tool targets a distinct operation: fetching a full article, searching articles, getting industry overviews, listing article metadata, and listing industries. No overlap in functionality.
All tool names follow a consistent verb_noun pattern using snake_case (get_article, search_articles, get_industry_overview, list_articles, list_industries).
With 5 tools, the set is well-scoped for a content publication server, covering listing and retrieval of both articles and industry overviews without excess.
The tool surface covers the full read lifecycle for articles (list, get, search) and industries (list, get overview). No obvious gaps for the intended purpose of accessing Kettle Logic content.
Unclaimed servers have limited discoverability.
If you are the server author, to access and configure the admin panel.
Read-only MCP server for Flamel.ai's public content: company overview, blog, case studies, FAQs.
Agent-native MCP server over the public saagarpatel.dev corpus. Read-only, stateless.
Read-only MCP server exposing a user ORANO library to their own AI agent.
Read-only MCP access to devplane.dev's public pages, FAQ corpus and comparisons.
We provide all the information about MCP servers via our MCP API.
curl -X GET 'https://glama.ai/api/mcp/v1/servers/mploschiavo/mcp-kettlelogic'
If you have feedback or need assistance with the MCP directory API, please join our Discord server
