Platform extensibility
Which AI agent platforms are extensible through APIs, SDKs and MCP?
Of 946 vendors, 520 document full extensibility: stable APIs, software development kits, consuming or serving Model Context Protocol tools, custom tool authoring and integration into internal systems. Another 180 document partial coverage and 246 document none.
Every vendor in the index is assessed against the same 14 point taxonomy from public documentation, and no vendor pays for placement. Counts on this page were measured across all 946 public vendors on September 30, 2026.
How the 946 vendors split
No public evidence means the reviewed sources did not document the capability. On this index that is a statement about the evidence, not proof that the capability is absent. See methodology.
What counts as full coverage
Full coverage means an outside caller can drive this platform through a documented API or SDK for the vendor's own product. A gated developer portal still clears Full when a readable page names the API, states what it covers and documents how to get access. The vendor's own MCP server clears Full without an API when it publishes its endpoint, its auth scheme and its tools, and those tools create or change the platform's own core objects, as does an A2A agent card with its endpoint and auth scheme documented. Partial means a CLI or a marketplace without an API or SDK, an MCP server that only reads or that the vendor presents as a sample, a read only API or webhook, or an API asserted with no reference behind it. Direction is the whole test: anything that lets outside code call this platform counts, and anything that lets this platform reach into other systems, including a bridge exposing the buyer's own tools to agents, belongs on integrations. One limit worth stating plainly: this axis bundles APIs, SDKs and MCP into a single grade, so the Full count is not a count of vendors supporting MCP and should never be quoted as one. Confirm MCP specifically with any vendor on your shortlist.
How to read these numbers
Agent infrastructure leads at 81 percent, which is close to a tautology since composability is the product, and the browser lane at 61 percent is the surprise, driven by the number of vendors now shipping an MCP server so that other people's agents can drive a browser. Healthcare at 14 percent is the floor. That lane sells into environments where integration goes through an interface engine and a compliance review rather than a public API, so the absence is structural rather than careless. If your plan is to assemble several vendors into one agent estate, this axis is the one that decides whether that plan is engineering or wishful thinking.
Leading platform for apis, sdks & mcp extensibility in each use case
Picked mechanically: the highest total coverage vendor in each lane that documents full evidence on this axis, one vendor per row. Scores are out of 14.
-
1. Appian, for enterprise operations agents
14 / 14Full enterprise operations agents ranking · Compare the whole lane on all 14 axes
-
2. FLOWX.AI, for multi-agent platforms
14 / 14Full multi-agent platforms ranking · Compare the whole lane on all 14 axes
-
3. Gumloop, for GTM and revenue agents
14 / 14Full GTM and revenue agents ranking · Compare the whole lane on all 14 axes
-
4. Mastra, for agent infrastructure platforms
14 / 14Full agent infrastructure platforms ranking · Compare the whole lane on all 14 axes
-
5. ServiceNow, for customer support agents
14 / 14Full customer support agents ranking · Compare the whole lane on all 14 axes
-
6. UiPath, for agent builders
14 / 14Full agent builders ranking · Compare the whole lane on all 14 axes
-
7. GitHub Copilot, for coding agents
13.5 / 14Full coding agents ranking · Compare the whole lane on all 14 axes
-
8. Dataiku, for data analyst agents
13 / 14Full data analyst agents ranking · Compare the whole lane on all 14 axes
-
9. HappyRobot, for voice agents
13 / 14Full voice agents ranking · Compare the whole lane on all 14 axes
-
10. Adopt AI, for browser and computer-use agents
12.5 / 14Full browser and computer-use agents ranking · Compare the whole lane on all 14 axes
-
11. Torq, for security and SOC agents
12.5 / 14Full security and SOC agents ranking · Compare the whole lane on all 14 axes
-
12. Edge Delta, for SRE and DevOps agents
12 / 14Full SRE and DevOps agents ranking · Compare the whole lane on all 14 axes
-
13. Innovaccer, for healthcare agents
11.5 / 14Full healthcare agents ranking · Compare the whole lane on all 14 axes
Documented coverage by use case
Share of each lane documenting full coverage on this axis. Vendors that sit in two lanes count in both, the same rule the rankings and matrices use.
| Use case | Full coverage | Share |
|---|---|---|
| agent infrastructure platforms | 156 of 186 | 84% |
| agent builders | 86 of 115 | 75% |
| coding agents | 48 of 64 | 75% |
| browser and computer-use agents | 31 of 42 | 74% |
| multi-agent platforms | 33 of 52 | 63% |
| SRE and DevOps agents | 22 of 37 | 59% |
| GTM and revenue agents | 70 of 138 | 51% |
| customer support agents | 54 of 113 | 48% |
| voice agents | 39 of 83 | 47% |
| data analyst agents | 27 of 59 | 46% |
| security and SOC agents | 36 of 85 | 42% |
| enterprise operations agents | 114 of 297 | 38% |
| healthcare agents | 21 of 69 | 30% |
Recent verified changes from the vendors named above
Capability coverage is not a static picture. These are the most recent sourced change log entries for the platforms listed above, newest first, one per vendor. Scores on this page update as entries like these are verified.
-
Mastra mcp / tool calling / api
High impactMastra released @mastra/mcp 2.0.0, whose servers support only the July 28, 2026 MCP revision and require @mastra/core 1.68 or newer. Tools that need caller input now suspend and resume through input_required continuations with signed request state. The release removes legacy initialization, session headers, standalone HTTP plus SSE transport, and earlier elicitation APIs.
September 22, 2026 · Verified · All Mastra changes
-
GitHub Copilot mcp / tool calling / api
High impactGitHub released the Copilot SDK for Java, enabling server-side Java applications to create agent sessions, register tools, and stream structured responses. The framework-agnostic SDK includes Bring Your Own Key (BYOK) support and integrates with Jakarta EE and Spring.
August 10, 2026 · Verified · All GitHub Copilot changes
-
ServiceNow mcp / tool calling / api
High impactServiceNow published an overview of Action Fabric, an interoperability framework connecting AI agents across the enterprise. It features an MCP Server Console for external clients to invoke ServiceNow tools, an MCP Client for ServiceNow agents to call external systems, and Agent2Agent protocol support for peer collaboration.
July 31, 2026 · Verified · All ServiceNow changes
-
UiPath mcp / tool calling / api
High impactUiPath launched a public preview of UiPath for Coding Agents, enabling developers to build automations directly from AI coding agents like Claude Code, Cursor, and Codex CLI. This update introduces a new command-line interface (uip CLI) and UiPath skills that allow coding agents to interact with a user's UiPath organization using natural language. It also introduces a Coding Agents for Test feature that allows these agents to autonomously read test cases, write execution results, and update requirement coverage data.
July 8, 2026 · Verified · All UiPath changes
-
Gumloop mcp / tool calling / api
Medium impactLarge MCP toolsets now lazy-load tool details only when needed, improving context efficiency and reducing unnecessary token usage.
June 9, 2026 · Verified · All Gumloop changes
Full change log · updated weekly across the whole index
Questions buyers ask
Does full coverage mean the vendor ships an MCP server?
Not necessarily. MCP is one form of evidence alongside a documented API, an SDK or custom tool authoring. Where MCP specifically is the requirement, confirm whether the vendor consumes MCP tools, serves them, or both.
Why does MCP matter for buyers?
It is becoming the common connector between agents and tools, so a platform that speaks it can be wired into an estate you already run rather than requiring a bespoke integration for every pairing.
Is an API enough on its own?
For most integration work, yes. For building agents that extend the platform itself, look for documented custom tools or an SDK, since a management API often only lets you configure what already exists.
The other 13 axes
No single axis decides a shortlist. Buyers who care about this one usually check model flexibility & routing and testing, debugging & optimization next, or open the full taxonomy to see how the 14 axes fit together.