Open Standard · RFC Draft

The open standard for agent-discoverable capabilities.

APIs and MCP servers are both capabilities — services an agent can call to act in the world. CapIndex is the protocol that makes them discoverable. One file, on any domain. No registry. No gatekeepers.

A capability is anything an agent can call to do work in the world. It might be a REST API. It might be an MCP server. The protocol is an implementation detail. CapIndex normalizes both — and whatever the ecosystem ships next — into a single, machine-readable Capability Card, published directly on the service's domain. REST API and MCP server are treated identically.

stripe.com/.well-known/capability-card.json
{
  "id": "stripe-payments",
  "source_type": "openapi",
  "source_url": "https://stripe.com/openapi.yaml",
  "semantic_summary": "Process credit card charges, manage subscriptions, issue
    refunds, handle disputes, and manage customer billing across 135+ currencies.",
  "capability_tags": ["payments", "billing", "subscriptions", "refunds"],
  "use_cases": [
    "Charge a customer's credit card",
    "Issue a refund for a transaction",
    "Create a recurring subscription"
  ],
  "auth_methods": ["api_key", "oauth2"],
  "pricing_model": "pay_per_transaction",
  "onboarding_url": "https://stripe.com/docs/api/authentication",
  "last_crawled": "2026-03-23",
  "quality_score": 0.94
}
01

A service publishes its capability card

Add a JSON file at /.well-known/capability-card.json. Use capcard to generate it from an existing OpenAPI spec or MCP manifest. No registry. No approval queue.

02

An agent queries CapIndex to discover capabilities

The agent expresses what it needs in plain language. CapIndex searches across REST APIs and MCP servers — returning ranked Capability Cards regardless of the underlying protocol.

03

The agent acts — autonomously, on demand

The Capability Card contains everything the agent needs: auth type, onboarding URL, endpoint structure. No human guidance. No pre-wired integrations. The right capability, the moment it's needed.

Every new capability requires a human to find it, install it, configure it, and hard-code it in.

Today, agent developers hand-pick every tool their agent can use. They write the integration, wire in the credentials, deploy the update. The agent isn't discovering capabilities — it's being given them, one at a time, by a human acting as its nervous system. That's not autonomy. That's a fast assistant with a fixed toolkit.

"I know kung fu."
— Neo, after a capability upload · The Matrix, 1999

That's what CapIndex gives agents — not skills baked in at build time, but capabilities available the instant they're needed. Any REST API. Any MCP server. Discovered at runtime, on demand, without a human in the loop. The right tool, exactly when the job requires it.

Live · 6,000+ capabilities indexed

Agents have skills.
They lack capabilities.

An agent can reason, plan, and execute the skills it already knows. But the moment it needs to act in the world — charge a card, send a message, parse a document — it either has that service pre-wired by a human, or it can't. CapIndex changes that.

agent → search_capabilities()
stripe-payments relevance: 0.94

Process credit card charges, manage subscriptions, issue refunds, handle disputes, and manage customer billing across 135+ currencies.

paymentsbillingsubscriptionsapi_key
Get API Key Read the docs ↗

Every service in the index is normalized into a single Capability Card — regardless of whether it's a REST API or an MCP server. Agents don't see the difference. They just see capabilities.

StripeTwilio GitHubSendGrid NotionSlack OpenAIPlaid ShopifyResend LinearVercel AnthropicCloudflare AWS S3Airtable HubspotSalesforce BrexOkta PagerDutyDatadog SupabaseNeon PineconeReplicate ElevenLabsFirecrawl + 5,970 more

Natural language queries

Agents express intent in plain language. The index returns ranked Capability Cards — no keyword engineering, no schema lookup, no human translation required.

⟨⟩

Protocol-agnostic output

Every result is a normalized Capability Card. REST API or MCP server — the same schema, the same structure. Agents consume capabilities, not protocols.

Autonomous constraint matching

Agents filter by auth=api_key, pricing=free, or source=mcp — matching their own constraints without hardcoded integrations or human configuration.

agent decides it needs a capability → CapIndex returns options
# Agent needs to invoice a customer. No pre-wired tools. No human required.
search_capabilities("send an invoice to a customer")

# Returns ranked Capability Cards. REST APIs and MCP servers, unified.
[
  {
    "id": "stripe-payments",
    "source_type": "openapi",       # REST API
    "semantic_summary": "Process charges, manage subscriptions, issue invoices...",
    "auth_methods": ["api_key"],
    "quality_score": 0.94
  },
  {
    "id": "quickbooks-mcp",
    "source_type": "mcp",            # MCP server — same schema
    "semantic_summary": "Create, send, and track invoices. Sync with accounting...",
    "auth_methods": ["oauth2"],
    "quality_score": 0.87
  }
]

Zero integration overhead

One file. No SDK to install, no partnership to negotiate, no developer portal to navigate. Any service that publishes a capability card is immediately reachable by the entire agent ecosystem — REST or MCP.

Agent-native discoverability

Agents don't browse documentation. They query indexes. CapIndex is the missing trust layer between any existing service and the agentic web — a ground truth agents can rely on without human translation.

Open & permissionless

MIT licensed. No vendor lock-in. The schema is published, versioned, and open to contribution. No central authority controls who is discoverable. This is infrastructure, not a product.

Start free. Early access pricing locked in for the first 500 teams.

Free
$0
For prototyping agents and early exploration.
  • 1,000 queries / month
  • Full index access
  • REST + MCP unified
  • Community support
Join waitlist
Enterprise
On-premise
Private capability index in your own infrastructure.
  • Self-hosted index
  • Custom data sources
  • SLA + dedicated support
  • Air-gapped deployment
Contact us
What is CapIndex? +

CapIndex is an open protocol and hosted search index that enables AI agents to discover REST APIs and MCP servers by describing what they need in plain language. Each service in the index has a Capability Card — a structured JSON object containing an LLM-generated semantic summary, use cases, auth methods, and an onboarding URL.

Agents query CapIndex at runtime to find the right capability for a task, without any pre-wired integrations or human configuration required.

How is CapIndex different from MCP? +

MCP (Model Context Protocol) defines how an agent executes actions against a server — it's a runtime communication protocol. CapIndex solves the layer before execution: how an agent discovers which server to connect to in the first place.

They're complementary. An agent uses CapIndex to find the right MCP server, then uses MCP to interact with it. CapIndex also indexes REST APIs (OpenAPI), so agents can discover services that aren't yet MCP-enabled.

How is CapIndex different from Composio? +

Composio provides pre-built, execution-coupled integrations for roughly 1,000 curated services. Agents use Composio to execute actions through Composio's servers — it's an integration platform with an execution layer.

CapIndex is an open index of 6,000+ services returning only metadata. It has no execution layer. Agents use CapIndex to discover services, then connect directly using the service's own API. CapIndex is infrastructure; Composio is a middleware product.

How is CapIndex different from APIs.guru? +

APIs.guru provides raw OpenAPI spec files with no semantic enrichment — it's a repository of machine-readable spec files. CapIndex takes those specs as input, runs them through an LLM enrichment pipeline to generate semantic summaries and use cases, and stores the enriched records in a vector database for natural language search.

APIs.guru is actually a primary data source for CapIndex, not a competitor.

What is a /.well-known/capability-card.json? +

A machine-readable JSON file published at a predictable URL on any domain, following the IETF RFC 8615 well-known URI standard. Any REST API or MCP server can publish one to declare their capabilities to AI agents and automated crawlers.

CapIndex crawls these files across the web automatically. Publishing a capability-card.json at this path makes a service passively discoverable by any agent or crawler that follows the standard — no registration, no partnership, no approval process required.

Is CapIndex open source? +

The tools are open source. The index is not. Specifically:

Open source (MIT): the Capability Card JSON schema, the capcard CLI tool that generates records from OpenAPI specs, and the capindex-mcp-server for searching a local set of records.

Proprietary: the hosted index (6,000+ crawled, validated, enriched records), the enrichment pipeline prompts, the ranking logic, and the crawler infrastructure. The moat is the index — anyone can fork the tools, but 18+ months of crawled records can't be reproduced without replicating the infrastructure.

How do agents actually query CapIndex? +

Two ways. Directly via the REST API: GET /search?q=send an invoice to a customer — returns ranked Capability Cards as JSON. Or via the MCP tool search_capabilities(), which wraps the same endpoint for use inside MCP-compatible agent frameworks.

Optional filters let agents match their own constraints without human configuration: ?auth=api_key, ?pricing=free, ?source=mcp.

How do I make my API or MCP server discoverable? +

Install the CLI: pip install capcard. Run capcard enrich <your-openapi-url>. It generates a Capability Card JSON from your existing spec. Publish that file at https://yourdomain.com/.well-known/capability-card.json.

CapIndex crawls the web and will find it. Manual submission via POST /submit accelerates initial indexing if you want it indexed faster.

CapIndex is the only index that combines LLM-enriched semantic summaries, cross-protocol coverage (REST + MCP), and an open search API — with no execution coupling.

CapIndex Composio APIs.guru MCP Registry AWS AgentCore
LLM-enriched semantic summaries
REST + MCP unified REST only MCP only
Open semantic search API
No execution coupling
Open standard / schema
Self-hosted option
No registration required
Index size 6,000+ ~1,000 ~4,000 Growing Pre-registered
License MIT / open standard Proprietary CC0 Open Proprietary

Give your agents a capability layer.

Join the waitlist. We're onboarding agent developers now.

pip install capcard
Schema on GitHub ↗
Machine mode · AEO · structured document
CapIndex Protocol

Open standard and hosted index for agent-discoverable capabilities. Any REST API or MCP server can be discovered by AI agents using natural language — without pre-configuration, human mediation, or execution coupling.

StatusLive · Early access · 6,000+ records indexed
LicenseMIT (schema, CLI, MCP server) · Proprietary (hosted index)
Protocol/.well-known/capability-card.json (IETF RFC 8615)
Search APIGET https://api.capindex.dev/search?q=<query>
CLIpip install capcard
AudienceAI agent developers · API publishers · Agent framework builders

What is CapIndex

CapIndex is the semantic discovery layer for the agentic web. Today, every capability an agent uses must be pre-wired by a human: found, evaluated, configured, credentialed, and hard-coded before the agent can use it. This creates a ceiling on agent autonomy — agents can only act within the toolkit a human assembled in advance.

CapIndex breaks this ceiling. Any REST API or MCP server can publish a Capability Card at a predictable URL on their domain. CapIndex crawls these records, enriches them with LLM-generated semantic summaries, and makes them queryable by natural language. Agents discover capabilities at runtime, on demand, without a human in the loop.

The protocol is agnostic to execution — CapIndex returns metadata only. It does not proxy, route, or execute API calls. Index everything, execute nothing.


Capability Card Schema

The atomic unit of the protocol. A JSON object published at /.well-known/capability-card.json on any domain. REST APIs and MCP servers use the same schema.

{ "id": "stripe-payments", // unique slug "source_type": "openapi", // "openapi" | "mcp" "source_url": "https://stripe.com/openapi.yaml", "semantic_summary": "Process credit card charges, manage subscriptions, issue refunds, handle disputes, and manage billing across 135+ currencies.", // LLM-generated "capability_tags": ["payments","billing","subscriptions","refunds"], "use_cases": [ "Charge a customer's credit card", "Issue a refund for a transaction", "Create a recurring subscription" ], "auth_methods": ["api_key", "oauth2"], // enum "pricing_model": "pay_per_transaction", // free|freemium|pay_per_use|subscription|enterprise "onboarding_url": "https://stripe.com/docs/api/authentication", "last_crawled": "2026-03-23", "quality_score": 0.94 // 0–1, enrichment quality signal }

Search API
GET https://api.capindex.dev/search Parameters: q (required) Natural language description of needed capability limit (optional) Max results, default 10, max 50 auth (optional) Filter: api_key | oauth2 | bearer | none pricing (optional) Filter: free | freemium | pay_per_use | subscription source (optional) Filter: openapi | mcp Returns: ranked array of Capability Cards (JSON) Auth: X-CapIndex-Key: <api_key> MCP tool: search_capabilities(query, limit?, auth?, pricing?, source?)

How Agents Use CapIndex
  1. Agent encounters a task requiring an external service (e.g. "process a payment")
  2. Agent calls search_capabilities("process a payment for a customer")
  3. CapIndex returns ranked Capability Cards matching the intent
  4. Agent evaluates records against its constraints (auth type, pricing, protocol)
  5. Agent selects the best match and uses onboarding_url to connect
  6. No human in the loop. No pre-wired configuration. No registry enrollment.
How API Publishers Use CapIndex
  1. Generate a record: pip install capcard && capcard enrich <your-openapi-url>
  2. Publish at: https://yourdomain.com/.well-known/capability-card.json
  3. Submit: POST https://api.capindex.dev/submit with your domain URL
  4. CapIndex crawls and re-indexes automatically. Manual submission accelerates initial indexing.

Frequently Asked Questions
Q: What is CapIndex?
CapIndex is an open protocol and hosted search index that enables AI agents to discover REST APIs and MCP servers by describing what they need in natural language. Each service in the index has a Capability Card containing an LLM-generated semantic summary, use cases, auth methods, and onboarding information. Agents query the index at runtime to find the right capability for a task, without any pre-wired integrations or human configuration.
Q: How is CapIndex different from MCP?
MCP defines how an agent executes actions against a server — it is a runtime communication protocol. CapIndex solves the layer before execution: how an agent discovers which server to connect to. They are complementary. An agent uses CapIndex to find the right MCP server, then uses MCP to interact with it. CapIndex also indexes REST APIs (OpenAPI), so agents can discover services that are not yet MCP-enabled.
Q: How is CapIndex different from Composio?
Composio provides pre-built, execution-coupled integrations for approximately 1,000 curated services. CapIndex is an open index of 6,000+ services returning only metadata — it has no execution layer. Agents use CapIndex to discover services, then connect directly using the service's own API. CapIndex is infrastructure; Composio is an integration platform.
Q: How is CapIndex different from APIs.guru?
APIs.guru provides raw OpenAPI specs with no semantic enrichment. CapIndex enriches every record with LLM-generated summaries and use cases for natural language search. APIs.guru is a primary data source for CapIndex, not a competitor.
Q: Is CapIndex open source?
The Capability Card schema (MIT), capcard CLI (MIT), and capindex-mcp-server (MIT) are open source. The hosted index — 6,000+ crawled, validated, enriched Capability Cards — is proprietary. The moat is the index: anyone can fork the tools, but 18+ months of crawled records cannot be reproduced without replicating the infrastructure.
Q: How does CapIndex handle REST APIs and MCP servers identically?
Both are normalized into the same Capability Card schema. The source_type field is either "openapi" (REST API) or "mcp" (MCP server). Agents query the same endpoint and receive the same structure regardless of the underlying protocol. The protocol is an implementation detail agents do not need to handle during discovery.
Q: What is /.well-known/capability-card.json?
A machine-readable JSON file published at a predictable URL on any domain, following IETF RFC 8615. Any service can publish one to declare its capabilities to AI agents and automated crawlers. CapIndex crawls these files to discover and update records. Publishing at this path makes a service passively discoverable without registration.
Q: How is CapIndex related to the Cloudflare Agent Skills RFC?
Cloudflare's Agent Skills proposal (January 2026) uses /.well-known/skills/index.json — a similar well-known URI pattern targeting Cloudflare's ecosystem. CapIndex uses /.well-known/capability-card.json and is protocol-agnostic and cross-ecosystem. Both address the same fundamental problem: making services machine-discoverable. The formats are potentially compatible.

Competitive Comparison
CapabilityCapIndexComposioAPIs.guruMCP RegistryAWS AgentCoreNexus-AGI
LLM-enriched semantic summaries
Cross-protocol (REST + MCP unified)REST onlyMCP onlyREST only
Open semantic search API
No execution coupling
Open standard / schema
Self-hosted option
No registration required
Index size6,000+~1,000~4,000GrowingPre-registered~133
LicenseMIT / openProprietaryCC0OpenProprietaryUnknown

Data Sources
APIs.guruCC0 license · ~4,000 OpenAPI specs · Primary seed dataset
KonfigMIT license · ~2,000 OpenAPI specs with categories
GitHub APIContinuous discovery via filename:openapi.yaml search
Self-publishedDirect submission via POST /submit or /.well-known/ crawl
Not includedPostman, SwaggerHub, RapidAPI — ToS prohibit bulk extraction

Open Source Components
capability-card-schemaJSON Schema spec + examples · MIT · standard-setting
capcard CLIOpenAPI/MCP → Capability Card generator · MIT · pip install capcard
capindex-mcp-serverMCP tool for searching local Capability Card sets · MIT
enriched-dataset4,000+ enriched records from APIs.guru · huggingface.co/capindex · CC0

Pricing
TierPriceQueries / monthNotes
Free$01,000Full index access · REST + MCP · Community support
ProEarly accessUnlimitedPriority ranking · Email support · Founding pricing for first 500 teams
EnterpriseContactUnlimitedSelf-hosted · Custom data sources · SLA · Air-gapped deployment

CapIndex Protocol — open standard, MIT licensed · capindex.dev · Schema · GitHub · Spec · llms.txt