Endpoint
| Transport | Streamable HTTP (MCP 2025-11-25) |
|---|---|
| URL | https://enhanciar.in/enhanciar/mcp/ |
| Auth | Header Authorization: Bearer dh_<your key> |
| Workspace | Chosen when you create the key. One key = one workspace; add a second server for a second workspace. |
Set-up
Claude Code
claude mcp add --transport http --scope user enhanciar \
https://enhanciar.in/enhanciar/mcp/ \
--header "Authorization: Bearer dh_your_key"
Cursor
Settings → Tools & Integrations → New MCP server → type HTTP, URL https://enhanciar.in/enhanciar/mcp/, header Authorization: Bearer dh_your_key.
Claude Desktop
Claude Desktop’s Connectors UI cannot send a custom header, so bridge it with mcp-remote in claude_desktop_config.json:
{
"mcpServers": {
"enhanciar": {
"command": "npx",
"args": ["-y", "mcp-remote", "https://enhanciar.in/enhanciar/mcp/",
"--header", "Authorization: Bearer dh_your_key"]
}
}
}
Settings → Connections in the app has copy-paste snippets for each client with your key filled in.
Tools
| Tool | What it does |
|---|---|
query | Ask a question in plain language; returns a synthesised answer with a citation on every claim (file:line, Slack permalink, doc paragraph). |
search_wiki | Search the generated wiki. |
get_page / list_pages | Read one wiki page in full; list every page the key can see. |
impact | Blast radius of changing a file, function or module: callers, callees, affected tests and pages, from the graph. |
get_graph | The entity / community graph manifest. |
get_process_map | How a process actually runs across code, people and tools. |
list_repos | Repositories the workspace has ingested. |
list_skills / get_skill | Reusable skills distilled from the codebase. |
propose_action / list_proposed_actions | Draft a Jira, Linear, GitHub, Slack or calendar action into the review queue, and list what is waiting. The server can propose; only a person in the app can approve or execute. |
The review queue is the same Actions queue your team sees in the app, and the pull-request reviews Enhanciar posts come from the same brain, so what an agent proposes, what a person approves and what a reviewer reads all share one source of truth.
What it can never do
- Execute an action. Proposals wait for human approval in the app.
- Reach beyond you. A key is tied to you and one workspace; leave the workspace and the key stops on the next request.
- Bill you for tokens. Model calls run under your own key (BYOK); Enhanciar adds no markup.
A worked example
Say you are in Cursor, about to change how checkout rate limiting works, and you have connected the Enhanciar server.
- You ask the agent: “Why is checkout limited to 20 requests a second?” It calls
query. The answer cites the file and line where the limit lives, the pull request that added it and the Slack thread where it was decided, so you can open each one and check. - You ask: “What breaks if I change it?” It calls
impactand lists the callers, the tests and the wiki pages that depend on that code. - You ask it to track the follow-up. It calls
propose_actionto draft a GitHub or Jira issue. The draft lands in the Actions queue in the app; nothing is created until a person on your team opens it there and clicks Approve.
The agent never sees more than the workspace the key belongs to, and every model call runs on your own AI key.
Common questions
Do I need an account?
Yes. API keys are created in the app under Settings → API keys, so you need an Enhanciar account. Each key is tied to you and one workspace.
Which clients work?
Any client that speaks streamable HTTP MCP and can send an Authorization header: Claude Code and Cursor directly, Claude Desktop through mcp-remote as shown above.
What is the server URL?
https://enhanciar.in/enhanciar/mcp/ over streamable HTTP, with the header Authorization: Bearer followed by your key.
Can an agent create tickets or send messages through it?
No. The propose_action tool only drafts a Jira, Linear, GitHub, Slack or calendar action into the review queue. A person on your team approves it in the app before anything runs.
Does the MCP server cost extra?
No. It comes with every paid plan (Solo $24 a month, Team $100 a month for up to 10 people). Model calls run on your own AI key, and Enhanciar adds no markup.
Why a remote server and not a plugin
Every MCP client already knows how to talk to a remote server, so there is nothing to install or keep updated per editor. The registry entry, the health checks on Glama and the tool schemas are public, so you can inspect exactly what your agent gets before you connect. Questions: enhanciar@gmail.com.