How MCP works
The protocol underneath all of this: tool discovery, authorization, credentials, and the parts of the specification that are still moving.
Connect AI agents to an internal API, no server
How to connect AI agents to an internal API by authoring a custom connector from config, and when writing your own MCP server is still better.
Roopendra Talekar
When to fork public MCP connector definitions, and when not to
Three cases to fork public MCP connector definitions in Elaichi, one case to file the change upstream, and how pull review sorts the differences.
Roopendra Talekar
JSON-RPC 2.0 over Streamable HTTP MCP: one request on the wire
Trace one MCP tool call end to end: the JSON-RPC 2.0 over Streamable HTTP MCP envelope, the stateless POST, and what OAuth adds on top.
Raajshekhar Rajan
Roles, Restrictions and ACLs in MCP: Three Layers
Roles restrictions and ACLs in MCP are three layers: a role says what a member may do, a restriction which tools they reach, an ACL who may use one connection.
Resource ACL view use edit: how Elaichi shares access
How the resource ACL view use edit grant works in Elaichi, and why no org-level permission widens what a member can list.
Uday Gajavalli
MCP connector lineage identity and fork blocks
MCP connector lineage identity is a connector's own slug plus the forks that declare descent from it, walked toward the root, block-only, and failing closed.
Uday Gajavalli
Where MCP restriction enforcement points actually sit
The mcp restriction enforcement points in Elaichi run at eight places, not four, and a ninth check inside the AI client would prove nothing.
Nachi Raman
Enterprise MCP: a buyer's guide for IT teams
Enterprise MCP for buyers: what changes when MCP goes from one laptop to a whole company, and what to ask a vendor before you sign.
Roopendra Talekar
The MCP Connector JSON Config Schema: What Saves and What Fails at Execution
How the mcp connector json config schema is validated, how secrets become secret_paths, and three mistakes that pass the save and fail at execution.
MCP execute_tool indirection security
MCP execute_tool indirection security rests on one rule: the wrapper unwraps to a name and arguments, then meets the gates a direct call would meet.
Raajshekhar Rajan
How MCP tool restriction enforcement works, end to end
Follow one rule through MCP tool restriction enforcement in Elaichi, from the admin who writes it to the final check on the resolved outbound URL.
Roopendra Talekar
MCP for coding agents across an engineering org
MCP for coding agents needs one endpoint, a sign-in per engineer, and rules the agent cannot edit, because nobody reads each call it makes.
Roopendra Talekar
What is an MCP control plane, and who needs one?
An MCP control plane is one org-wide MCP endpoint that serves the tools, signs each person in and decides access on every call.
Roopendra Talekar
MCP enterprise-managed authorization: what IT owns
MCP enterprise-managed authorization has been stable since June 2026, and Elaichi does not implement it. Here is what SAML, OIDC and SCIM decide for you.
Nachi Raman
How MCP tool search picks one tool from hundreds
MCP tool search in Elaichi scores tools by matching words, then returns nothing unless a tool covers at least half the query, with rare words weighted most.
Raajshekhar Rajan
The MCP context window problem, and a fix
The MCP context window problem: each listed tool sits in the model's prompt on every turn. Elaichi lists no connected tools; the model searches for them.
Uday Gajavalli
What is an MCP gateway? The four shapes
What is an MCP gateway: one address between AI clients and their tools that signs people in, applies rules and records calls. It comes in four shapes.
Nachi Raman
MCP server registry vs first-party connectors
MCP server registry vs first-party connectors comes down to who fixes a broken tool: each server's own author, or one vendor that wrote and serves them.
Roopendra Talekar
OAuth or API keys for AI agents?
Choosing OAuth or API keys for AI agents comes down to revocation: a grant is checked on every call, while a key works until someone rotates it.
Roopendra Talekar
Put AI to work in your own apps
14 days of Black, no credit card. Start with one app and one team.
When the trial ends, nothing is deleted. You move to the free Personal plan, and connections, roles and the audit log stay where they are.
Add one link to your AI app
The same link for everyone on your team.
https://api.elaichi.ai/mcp
ChatGPT
Claude
Cursor
Microsoft Copilot