The Merge Agent Handler alternatives that fit employee AI access are products built for a company's own staff, not for agents a product ships to customers. Elaichi is one: Claude, ChatGPT, Cursor or any MCP client signs in to one address, each call acts as the person who approved it, and the price is per seat. Composio, Arcade and Zapier MCP are the other names on the same shortlist, each set out below with its own page and the date it was read.
Merge's FAQ describes Agent Handler as "a platform for securely connecting AI agents to enterprise tools at scale". The same page documents SCIM-based provisioning, DLP scanning, policy guardrails, a searchable audit trail, and usage-based pricing that is free to start (merge.dev, checked October 2026). Elaichi vs Merge Agent Handler then turns on three forks: the address, what the AI client holds, and who writes the connectors. Where Merge's page does not settle a fork, this post puts it to Merge as a question.
The request usually arrives from one team. Support wants Claude to read Zendesk. Two weeks later finance wants ChatGPT near Xero, and legal has questions about both. You are now picking a shape you will have to defend in a year.
Which Merge Agent Handler alternatives fit employee AI access?
Four alternatives come up when the users are employees: Elaichi, Composio, Arcade and Zapier MCP. They differ on how many addresses exist, who each call acts as, and whether the bill counts seats or calls.
Elaichi is a governed MCP control plane for a company's own staff. MCP (Model Context Protocol) is the standard way an AI assistant calls tools in other apps. Every client points at one address, https://api.elaichi.ai/mcp, and each person approves their own grant. Admins restrict a whole connector or a single tool per role or per member, and each call that reaches execution is written to one audit log. Gold lists at $15 per user per month in USD, or $10 per user per month billed annually, with SSO and SCIM included.
Composio's MCP Gateway page says each team "gets its own MCP endpoint carrying only the tools it is permitted to use". Administrators can manage access by organization, team, user, agent or toolkit (Composio MCP Gateway, checked November 2026). Composio meters tool calls, with a $0 Hobby plan and a $29-a-month Pro plan (Composio pricing, checked November 2026).
Arcade gives each gateway its own URL, and its docs recommend sign-in through your OIDC identity provider for production (Arcade MCP Gateways, checked November 2026). Its Team plan is a $25 monthly platform fee plus $0.01 per tool call (Arcade pricing, checked November 2026).
Zapier MCP gives every member a server per client, and each successful tool call consumes two tasks from the Zapier plan (Zapier MCP usage, checked September 2026). It fits a company that already runs Zaps across its apps. Arcade, Merge and Composio side by side goes deeper on three of these, and the MCP gateway shortlist covers the wider field.
Elaichi vs Merge Agent Handler: what can be compared?
Every Elaichi row below can be stated from the product itself. A Merge row cites Merge's page where it answers, and is a question where it does not. This blog states a named vendor's behavior only from that vendor's own page, with the date it was read: for Agent Handler, Merge's product page (checked October 2026).
| Row | Elaichi | Merge Agent Handler |
|---|---|---|
| Address model (fork one) | One organization-wide POST /mcp endpoint behind OAuth; every AI client points at the same address |
No endpoint pattern published. Ask Merge: how many endpoints exist at 200 employees, and who creates each one |
| Credential shape (fork two) | The client holds an OAuth grant only; per-account secrets sit in a separate credential service, AES-256-GCM at rest | Ask Merge: what the client stores, and whether any part of it is long-lived |
| Connector authorship (fork three) | Elaichi serves 650+ connectors and authors most of them, the rest being vendors' own MCP servers; custom or forked connectors via JSON config, with a review surface for upstream changes | Ask Merge: who writes the connectors, and who fixes one when the upstream API changes |
| Audit | Keeps one entry for each connected-tool call that reaches execution (a call refused earlier writes none), naming the account reached; logs the one path argument that names the object, as the target id, and nothing else about the arguments; free read-only Auditor seat | Searchable audit trail documented. Ask Merge: which calls get an entry, and does it name the connected account reached |
| Offboarding | Grant revocation effective on the next call; a preflight lists every connection the departing member owns; a private one a shared toolbox relies on blocks removal until it is transferred to one member or deleted | Ask Merge: is removal effective by cache expiry or by transaction, and what happens to personal connections |
| Pricing | Per user: Gold lists at $15 per user per month in USD, or $10 per user per month billed annually (pricing); 14-day trial | Usage-based and free to start. Ask Merge for the rate |
Borrow the same rule for your own evaluation. Read each vendor's product documentation rather than a comparison grid, this one included. A vendor whose docs answer the address, credential and connector questions in plain language can be evaluated in an afternoon.
Is Merge Agent Handler worth it for employees?
Merge Agent Handler is worth it when the job matches what Merge describes: connecting AI agents to enterprise tools at scale, with SCIM provisioning, DLP and an audit trail for IT (merge.dev, checked October 2026). For employees in their own Claude, ChatGPT or Cursor, test three things first. Does each call act as the person? Can a rule name one tool? Can finance forecast the bill?
Any layer of this kind, Merge's or Elaichi's, is worth buying once more than one team reaches more than one app from more than one AI client. Below that line it is overhead with a monthly bill.
What you are buying is not connectivity. Claude, ChatGPT and Cursor already connect to app vendors' own MCP servers with nothing in between. A control plane adds four things: one address to hand out, one place to say who may run which tool, one record of what ran, and one action when somebody leaves. Those four start to pay for themselves at the point where a spreadsheet of per-person setups would otherwise start.
The cost side is easy to check before the demo. Elaichi has three plans: Personal, which is free, then Gold and Black. Gold lists at $15 per user per month in USD, or $10 per user per month billed annually, and every new workspace starts on a 14-day Black trial, no credit card to start. Buying Gold or Black during the trial keeps its remaining days rather than granting a fresh 14, so it is one continuous trial, and the seats are first charged when it ends. A workspace that buys nothing moves to the free Personal plan. The pricing page shows the price for your region. Count the seats you would actually govern rather than headcount, because Guest, Billing Admin and the read-only Auditor take no billable seat. Merge prices Agent Handler on usage and says it is free to start (merge.dev, checked October 2026), so ask Merge for the rate.
The honest floor: two people in one team using one app do not need either product. That case is worked through in when a gateway is premature.
How is Merge Agent Handler priced, and what do the alternatives cost?
Merge's FAQ says Agent Handler "is free to start. Pricing plans are usage-based", with Enterprise on custom pricing or user seats, and it says audit logs are available on all plans (merge.dev, checked November 2026). That page publishes no rate, so ask Merge for one before you model a bill. Elaichi prices the other way round. It has three plans: Personal, which is free, then Gold and Black, Gold lists at $15 per user per month in USD, or $10 per user per month billed annually, and the number of tool calls does not change the bill.
| Product | Billing unit | What the vendor publishes |
|---|---|---|
| Merge Agent Handler | Usage, or user seats at Enterprise | Free to start, usage-based plans, Enterprise custom pricing or user seats, audit logs on all plans (merge.dev, checked November 2026) |
| Composio | Tool calls | Hobby $0, Pro $29 a month (pricing, checked November 2026) |
| Arcade | Tool calls | Free with 2,000 calls a month; Team $25 a month plus $0.01 a call (pricing, checked November 2026) |
| Zapier MCP | Tasks from the Zapier plan | Two tasks per successful call (usage, checked October 2026) |
| Elaichi | Seats | Gold at $15 per user per month in USD, or $10 per user per month billed annually; Guest, Billing Admin and Auditor seats are free |
Internal employee usage has a known shape. The number of people is fixed in advance, and the number of tool calls is not. Finance usually wants a figure it can budget before the quarter starts.
Neither unit is always cheaper. Per seat costs more when a handful of people make a huge number of calls. Per call costs more when usage is steady and high, such as an engineer querying Jira dozens of times a day. Measure your calls per employee in a one-week pilot before signing either contract. MCP gateway pricing per user works through the arithmetic.
Which option fits which scenario?
Pick by who the agent acts as and where it runs. Four cases cover most of these decisions.
| Scenario | Fits best | Why |
|---|---|---|
| An employee assistant across Claude, ChatGPT and Cursor | Elaichi | One address every client signs in to, a grant per person, and restrictions written per role or per member. Gold lists at $15 per user per month in USD, or $10 per user per month billed annually, so the bill follows headcount, not call volume. |
| Agents your own product ships to customers | Evaluate Merge Agent Handler | Merge describes Agent Handler as a platform for securely connecting AI agents to enterprise tools at scale (merge.dev, checked October 2026), so this is the case to evaluate it for. It is free to start and priced on usage (merge.dev, checked November 2026). A per-seat plane priced for your own staff does not match a per-customer workload. |
| A developer project that needs per-user sessions in application code | Composio | Composio's docs describe SDK sessions for developers, and a session can be exposed through a hosted MCP endpoint (docs.composio.dev, checked November 2026). |
| A small pilot: two people, one app, one client | Neither yet | A control plane over one connection is overhead with a monthly bill. Measure on a free tier or a trial first: Merge is free to start (merge.dev, checked November 2026), and every new Elaichi workspace starts on a 14-day Black trial, no credit card to start. Buying Gold or Black during the trial keeps its remaining days rather than granting a fresh 14, so it is one continuous trial, and the seats are first charged when it ends. A workspace that buys nothing moves to the free Personal plan. |
Arcade and Zapier MCP sit on the same shortlist for the first row. Arcade meters tool calls, at the price in the table above. Zapier MCP bills in Zapier tasks instead, two per successful tool call, drawn from the Zapier plan.
Fork one: is the address per user or per organization?
Elaichi gives the whole organization one address. Every SaaS account is connected once, and its tools are served through one organization-wide MCP endpoint, POST /mcp, behind OAuth. It speaks standard MCP over Streamable HTTP and JSON-RPC 2.0. The MCP specification defines that transport as one HTTP POST per message to a single endpoint (MCP transports). There are no per-toolbox URLs and no embedded tokens. There is no MCP server to create, list or revoke per person, so what varies between members is the grant, not the address.
That shows up at rollout. An admin adds the URL once where the client allows it, and each member then connects and signs in with their own grant. In Claude Team or Enterprise an owner adds it under Organization settings > Connectors, while on Pro or Max each person adds it under Customize > Connectors (Anthropic's guide). Cursor's admin MCP allowlist is Enterprise only and "does not push it to users' machines", so each developer adds it in their own Cursor (Cursor's docs). ChatGPT's steps are in connecting Elaichi to ChatGPT. A fourth MCP client is the same shape, not a fresh project.
Other shapes multiply what sits behind the address. Zapier MCP gives every client the same endpoint, but "Each MCP client gets its own MCP server", so each member holds one server per client. For clients that use a connection token, Zapier says to "give each user their own server and token rather than sharing one" (how Zapier MCP connections work, checked October 2026). Composio's MCP Gateway page says "each team gets its own MCP endpoint carrying only the tools it is permitted to use" (Composio MCP Gateway, checked October 2026). Both are built for a different reader, and both multiply the things you track. The arithmetic of one organization server against one per member is worked through on its own. Ask Merge how many addresses exist at 200 employees, and who creates each one.
Fork two: what does the AI client actually hold?
With Elaichi, the client holds an OAuth grant and nothing else. OAuth is the sign-in handshake that gives an application limited access without handing over a password. The MCP authorization spec makes the server an OAuth 2.1 resource server, and it says access tokens "MUST NOT be included in the URI query string" (MCP authorization). Elaichi embeds no token in its URL, and no secret sits in a client config file.
Revocation follows from that. The grant's revoked_at field is re-read from the organization store on every single call, with no cache. In Elaichi, removing or suspending a member revokes every live grant in the same transaction as the membership change, so the next call is already blocked. Role and restriction changes work differently. They resolve through a short cache and take effect within about two minutes.
Connector credentials never live in Elaichi. Per-account secrets sit in a separate credential service, encrypted with AES-256-GCM at rest, and that service owns refresh. A failed refresh marks the connection needs_reauth rather than failing quietly. A connect URL is a one-time session carrying no token, which is why it is safe to return over MCP. An organization can also supply its own OAuth app per connector, gated on connector:manage rather than connection:manage.
The contrast to test is a long-lived string in a config file. For clients outside its supported list, and for code, Zapier MCP uses a connection token. Zapier's docs say that token is long-lived, tied to one server, and "grants whoever holds it the ability to run the server's tools" (how connections work, checked October 2026). Ask Merge what the AI client stores, and what happens to it when the holder leaves.
Fork three: who writes the connectors you depend on?
Elaichi serves 650+ connectors, and authors and maintains most of them on its own infrastructure. The rest are vendors' own MCP servers, each published by Elaichi staff rather than drawn from an open registry. Your company does not run MCP servers. When a connector Elaichi wrote breaks, there is one party to fix it. Reports about a native MCP connector's tools go to the app's vendor.
Custom work has a path. A connector can be authored from JSON config or forked from a public connector. A fork can pull upstream changes through a review surface that separates new tools, safe updates, config diffs, conflicts and upstream removals. Conflicts and destructive removals stay unchecked by default. The connector:create permission is flagged high trust, because a custom connector can be pointed at any destination.
The trade-off, stated plainly: if the app you need is not in the catalog, somebody has to author or fork it. A vendor that lists servers other people publish may show that app on its site today, and whether it works next month depends on whoever publishes it. Merge's page counts its tools in the thousands (merge.dev, checked October 2026). Ask Merge who writes them, and who fixes one when the upstream API changes. The managed versus run-it-yourself cost breakdown covers the third option.
Who decides which tools each person can use?
Admins do, through three layers Elaichi keeps separate on purpose. Permissions are 58 action strings grouped into roles, and a member holds exactly one role, enforced by a unique index. Sharing is one grant of view, use or edit on a resource to a user, a team or the whole organization. Members see their own resources and the ones shared with them, and nothing else. Restrictions decide which connectors and which individual tools a target may reach.
In Elaichi, restriction targets are role or user only; the organization default is the absence of a rule, which means allow everything. A rule aimed at a member can only narrow what their role allows, and never replaces or loosens a role rule. Within each layer, blocks always beat allows. One trap: the allowlist stage engages on the presence of an allow rule, not its contents, so an allow rule that names nothing denies everything. A block can catch a tool by its name or by the operation behind it. An allow counts only the operation, because a tool's displayed name is a label the governed party can edit. The name against operation rule explains the asymmetry.
Frozen parameters handle the narrower case where one argument must not move. A frozen key is removed from the schema the model is shown, and its value is merged over whatever the caller sends at execution. Passing the key cannot unfreeze it. In Elaichi, a freeze binds only calls made through its toolbox entry, so share the toolbox with the people it governs, not the connection itself.
Which tools does the model see?
Control-plane operations, such as managing members or connections, are listed individually. App tools are handled differently: connected tools are never listed one by one, however few there are. The model looks a tool up through search_tools and calls it through execute_tool, which is only a naming indirection. It unwraps to the same name and arguments and falls through the identical gates.
Ranking is lexical over tool name, description and connector label, with a floor on relevance. A tool must account for at least half of the query's weight, where rarer words weigh more. Without a floor a search always returns something, and something from an app you did not ask about is worse than nothing, because the model calls it. The reasoning behind the floor sets out the incident that produced it.
What does the audit trail record after a tool call?
Elaichi's trail holds one entry for each connected-tool call that reaches execution (a call refused earlier writes none). The account on each entry is the one the call actually reached, taken from the execution rather than the intent. The trail logs the one path argument that names the object, as the target id, and nothing else about the arguments. Each entry also records the surface, mcp, and the OAuth client, with Claude, ChatGPT and Cursor marked verified. In Elaichi, a call from an MCP client is recorded under the person who signed in, with surface mcp and the OAuth client named, so the client is known at the point of action instead of guessed later from a user agent.
Two error strings exist per failed call. The one returned to the caller is derived from the third party's response body and is never written anywhere else. The one written to the audit trail is never derived from the request or the response, because audit records are organization-visible and fanned out to whatever SIEM the customer configured. In Elaichi, export to your own Datadog comes with the Black plan. Splunk HEC and Microsoft Sentinel are accepted as destinations, but Elaichi delivers events only to Datadog. The trail is append-only and eventually consistent, so a row may take a moment to appear. The read-only Auditor seat is free, so a compliance reviewer does not consume a license.
One honest limit. The prompt-injection write gate lives in the Elaichi agent window and does not apply to a raw tool call. What holds on the endpoint is a role check per operation, the forbidden classification, output redaction, OAuth scope limits and an audit row for each call that reaches execution.
How fast does access end when somebody leaves?
On the next call. In Elaichi, removing or suspending a member revokes every live grant in the same transaction as the membership change, so the departing person's next call fails. The cleanup around it is the part that usually gets missed.
Elaichi runs an offboarding preflight. It lists every connection the departing member owns. A private connection pinned by a toolbox its owner shared blocks the removal until an admin transfers it to one other active member or deletes it, where a toolbox is a named set of configured tools. At offboarding, a private connection that nothing beyond the member depends on cannot be transferred and is deleted with them. Nothing transfers to a team, the organization or the admin running the removal. A shared connection stays in place unless the admin names an action. Deleting it takes an explicit request. Otherwise transfer it to a member who is staying, which leaves every grant on it as it was. Delegated toolbox entries surface as a non-blocking warning, and re-pinning the entry is the fix. The contractor offboarding sequence walks through a live case.
Where is data held, and how do people sign in?
Elaichi has three regions, eu, us and apac, chosen at organization creation and fixed afterwards. For EU and US, the organization's data store and the execution of its org-scoped requests and tool calls stay in that jurisdiction. APAC uses best-effort placement and is not a residency guarantee. Connections created since 2026-10-05 have credentials placed in the organization's region, by best-effort placement for APAC. Older connections stay where they were until reconnected, which copies them into the region. Credentials are encrypted at rest with AES-256-GCM. The audit trail is stored in one EU log instance for every region.
Sign-in covers Google, GitHub and Microsoft, email codes, TOTP MFA with single-use recovery codes, and passkeys. SAML and OIDC SSO are built in-house, with SCIM v2 for users and groups and group-to-role mapping. Onboarding has four paths: emailed single-use invite links with roles and teams pre-assigned, verified-domain join offers with a configurable default role, SCIM provisioning, and just-in-time SSO.
When is Merge Agent Handler still the right pick?
Merge Agent Handler is the better buy when the agent serves your own customers rather than your employees. That agent lives inside a product you run and reaches many customers' instances of many systems. The identity on each call is then a customer tenant, and a per-seat control plane priced for your own staff does not solve it. Elaichi is the wrong tool for that job.
Usage pricing is the second case. When traffic is genuinely variable and you want the bill to track it, a meter matches a workload you cannot forecast a quarter out. Merge prices Agent Handler that way (merge.dev, checked October 2026).
If your engineers need per-user sessions in application code, rather than employees working inside Claude, look at a developer-first product. Composio's homepage addresses developers building agents, and its docs describe an SDK with per-user sessions and managed auth (composio.dev, docs.composio.dev, checked October 2026). There is a side-by-side with Composio here.
If two people in one team need one app, buy nothing yet. The case for holding off on a gateway is real, and a control plane over a single connection is overhead with a monthly bill. When the time comes, Gold lists at $15 per user per month in USD, or $10 per user per month billed annually, and the pricing page shows the price for your region.
What should you ask Merge before you sign?
Write these down and send them. The answers are more useful than any comparison table, this one included.
- Is Agent Handler built for agents your product ships to customers, or for employees working in Claude, ChatGPT and Cursor?
- How many endpoints exist for a 200-person company, and who creates each one?
- What does the AI client store, and is any part of it long-lived?
- When a person is removed, how long until their next tool call fails, and is that a cache or a transaction?
- Who authors the connectors, and who fixes one when the upstream API changes?
- Which tool calls get an audit entry, and does it name the connected account reached?
- Can you choose a region that holds the data store, the credentials and the calls?
- Is SSO over SAML and OIDC included, with SCIM provisioning and group-to-role mapping?
- What is the usage rate, and what changes at Enterprise volume?
What should you test before you buy?
Test four things in a trial account, with one real person and one real app. Each one has a yes or no answer that a slide cannot give you.
- Does each person sign in as themselves? Connect the same app for two members and run one call from each. The record should name two people, not one shared service account. In Elaichi every member approves their own OAuth grant, and each call acts as the person who approved it.
- Can a restriction be written per role and per member? Block one tool for a role, then block a different tool for one person holding that role. In Elaichi, restriction targets are role or user only; the organization default is the absence of a rule, which means allow everything. A rule aimed at a member can only narrow what their role allows, and never replaces or loosens a role rule. Time the change too: a role or restriction change takes effect within about two minutes.
- What happens when someone is removed? Remove the test member and retry the call. In Elaichi, removing or suspending a member revokes every live grant in the same transaction as the membership change. Ask what becomes of the connections that person made, and whether anything blocks the removal until a human decides.
- What does the audit log record per call? Run one write, then open the row. In Elaichi the trail holds one entry for each connected-tool call that reaches execution (a call refused earlier writes none), and it logs the one path argument that names the object, as the target id, and nothing else about the arguments.
Run the four against each vendor in the same hour, with the same app. Differences that a feature list hides show up in minutes.
How can two people evaluate both in a week?
Pick one team and one app, then give it a week and two people: an IT owner and one person from that team. Support with Zendesk works. So does sales with Salesforce.
Run the same script against each vendor. The point is not to see whether a read call succeeds, because any working product manages that. The point is to see what breaks when a person leaves, and what the record shows afterwards.
What decides the choice in the end?
Three forks decide it: one address or many, a grant or a stored secret, and connectors from the vendor or from a registry. Everything else is a feature list that will have changed by the time you renew.
The other vendor comparisons run the same structure against named products. Check the connector catalog for the apps your teams actually use, and the team use cases page for what a first rollout looks like.