
PCC — Physical Capability Cloud MCP server
Which MCP servers are PCC — Physical Capability Cloud's own, from the official registry and from asking its own domain, whether they answer, and what they offer. A program checked each of these, nobody copied them from a product page.
read 2026-10-26
- Registry namenetwork.capability/physical-capability-cloudverified in the MCP registry
The registry name sits under capability.network. The registry only lets a publisher use that namespace after proving control of the domain through DNS or HTTP.
- Stateanswerschecked 2026-10-26
- Tools256
- Server calls itselfPhysical Capability Cloud
- Server version2.19.3+sha256.a93b6fab5a8a325daafb456917
- Protocol version2025-06-18
On 2026-10-26 the server answered initialize and tools/list without a sign-in. First answer in 79 ms.
- Whether its tools do the product's job has not been judged yet.
- Checked on 2 of the 7 days it takes to judge reliability.
Once connected
- For "Publish a machine shop's process as a purchasable capability", the likely first call is provision_api_key.
- Keep approval on for pay_ip_royalty, cancel_protocol_run, pcc_revoke_scope, send_diagnostics, send_support_message: they delete, send or pay without saying so.
- create_kernel, pay_ip_royalty, create_protocol, update_protocol, publish_protocol change data without a sign-in: anyone with the address can call them.
- Its definitions take about 39663 tokens of context: enable only the tools a task needs.
Said by rules from the measurements above. The likely first call is a model's pick from the tool list; read its description before relying on it.
- INITE MCP scoreafter a weeka score needs 7 daily checks: 2 so far
- Use3 GitHub stars
- Answeredmeasuring2 of the 7 daily checks it takes to print a share
- Median first answer203 ms
- Space in an agent's context≈ 39,663 tokensthe definitions of all its tools (256), which an agent carries on every turn
- Tool definitionsA · 99 of 100
- Typed results7 of 256 toolsdeclared outputSchema: results come typed, read once
- Parameters without a description: in 10 of 256
What an agent can do with it
- can do · Publish a machine shop's process as a purchasable capability
- not sure · Compose independent capabilities into a multi-step production workflow
- not sure · Replace one inspection provider with another without restructuring
Each of the product's use cases, put to Jev against this server's tool list: can an agent do it with these tools?
Worth a look
- Tools that delete, send or pay, marked as not destructive: pay_ip_royalty, cancel_protocol_run, pcc_revoke_scope, send_diagnostics, send_support_message
- Tools that change data, on a server that answers without a sign-in: create_kernel, pay_ip_royalty, create_protocol, update_protocol, publish_protocol
A program checks the server once a day: initialize and tools/list, no model. Tokens are the definitions' length at about four characters a token. The grade averages six checks of every tool's name, description, schema and annotations; it says how well the tools are described, not whether they work. The score weighs what was measured and nothing else: answering (25%), doing the product's job (25%), definitions (20%), context (10%), use (10%) and how readily a client connects (10%). A part that was not measured is left out, and a hidden instruction caps the score at 40. The state of vendors' MCP servers: method and numbers →
provision_api_keyCALL THIS FIRST. Provisions an API key for the operator. Returns a pcc_live_* key that must be included as 'Authorization: Bearer <key>' on all subsequent requests. Accepts either email OR walletAddress. The key is shown once — save it…
list_api_keysList all active API keys for the authenticated operator. Returns key IDs, prefixes, scopes, rate limits, usage counts, and creation/expiry timestamps. Requires a valid API key or SIWE session.
revoke_api_keyRevoke an API key permanently. The key will immediately stop working. You must own the key. Use list_api_keys to find the key ID.
setup_detectAuto-detect current setup state: env vars, database, adapters, chain connectivity, storage, identity. Use this to see what's configured and what's missing before onboarding.
setup_generate_configGenerate a KERNEL_CONFIG JSON from device descriptions. Tell it what machines you have and it produces the config ready to paste into your environment.
setup_validateValidate a kernel config (20+ checks: JSON structure, device connectivity, adapter compatibility). Returns validation errors and warnings.
setup_register_deviceRegister a device in the database with adapter config and capabilities. Part of Step 3 in the onboarding flow. Required: kernelId, deviceId, type and adapterType (400 missing_required_fields otherwise).
setup_test_jobSubmit a test job to verify the full pipeline works end-to-end (Step 4 of onboarding). Returns job ID and result.
setup_statusComprehensive setup status across 6 categories: gateway, database, adapters, chain, storage, identity. Use to confirm everything is green before going live.
list_capability_typesList all capability types registered on the network (FDM, SLA, CNC, HPLC, etc.) with metadata.
search_capabilitiesSearch capability templates by type or keyword. Returns templates with pricing, assurance tiers, and availability.
get_build_optionsGet configuration options for a capability type (materials, dimensions, tolerances, assurance tiers). Use before calling calculate_price.
calculate_priceCalculate price for a capability configuration. Returns base price, tier premiums, and estimated total cost.
build_contractBuild and submit a capability contract with on-chain milestone escrow. Returns job ID and escrow address.
list_kernelsList all Shop Kernels on the network with status and capability types. Optionally filter by status.
get_kernelGet kernel details including full capability objects, devices, and queue.
get_kernel_devicesList all devices registered under a specific kernel, including adapter configs and device types.
get_kernel_jobsList all jobs submitted to a specific kernel, including job status and progress.
create_kernelRegister a new Shop Kernel (physical site) on the PCC network. Part of Step 2 in the onboarding flow.
list_jobsList all jobs with status. Optionally filter by kernelId or status.
get_jobGet job details including progress, evidence bundles, and milestones.
update_job_statusUpdate a job's status and optional progress percentage. Used by kernels to report job progress.
list_escrowsList escrow contracts with milestones and bonds. Optionally filter by status.
get_escrowGet escrow details. Pass a DB ID for database record (with milestones and disputes), or an EVM address (0x...) to read directly from the blockchain.
get_escrow_chain_stateRead full on-chain escrow state with all milestone details from the blockchain. More detailed than get_escrow for on-chain contracts.
get_escrow_eventsGet on-chain event history for an escrow contract. Returns funding, release, dispute, and bond events.
get_token_balanceRead an ERC-20 token balance for an account address from the blockchain.
get_token_allowanceRead the ERC-20 token allowance granted by an owner to a spender from the blockchain.
get_escrow_disputeRead dispute state for a specific milestone from the blockchain.
get_write_statusCheck if on-chain write operations are enabled (requires PCC_GATEWAY_PRIVATE_KEY to be configured). Returns write status and signer address.
fund_escrowFund a protocol-created escrow on-chain. Approval is handled INTERNALLY by the gateway for protocol-created escrows (audit C-03 removed the public approve step) — do NOT call a separate approve first; fund directly. Returns transaction…
release_milestoneRelease a milestone payment on-chain after the challenge window has expired. Returns transaction hash.
file_escrow_disputeFile a dispute against a milestone on-chain. Challenger must post a bond and provide evidence hash. Returns transaction hash.
deposit_bondDeposit an operator bond for a milestone on-chain. Required for Assurance Tier 3 jobs. Returns transaction hash.
submit_evidence_hashSubmit an evidence bundle hash for a milestone on-chain. Links the physical evidence to the on-chain settlement record. Returns transaction hash.
submit_attestationSubmit a verifier attestation for a milestone on-chain. Used by third-party verifiers (Bittensor subnet) to confirm evidence quality. Returns transaction hash.
archive_evidenceArchive an evidence bundle to IPFS/Storacha for permanent decentralized storage. Returns the content-addressed CID.
get_evidence_bundleGet an encrypted evidence bundle by its bundle ID. Returns the encrypted payload and key capsules.
grant_evidence_accessGrant a new recipient access to an encrypted evidence bundle. Creates a re-encrypted key capsule for the recipient.
list_evidence_grantsList all evidence access grants for an address. Returns bundles the address has been granted access to.
archive_encrypted_bundleArchive an encrypted evidence bundle to IPFS and store the resulting CID in the database. Idempotent — returns existing CID if already archived.
get_bundle_ipfsGet IPFS CIDs for a bundle — returns the primary CID, metadata CID, and Filecoin deal ID if available.
retrieve_ipfsRetrieve raw data from IPFS by CID. Used to fetch archived evidence bundles from decentralized storage.
get_lit_conditionsGet Lit Protocol access conditions for a Lit-encrypted evidence bundle. Returns the conditions that gate decryption.
lit_decryptDecrypt a Lit Protocol-encrypted evidence bundle using a Lit auth signature. Returns the decrypted bundle payload.
verify_evidence_zkVerify a ZK proof by its proof ID. Checks the proof against its verification key and public inputs.
commit_evidenceCreate a Merkle commitment for an evidence bundle hash. Used as the first step in ZK proof generation for Tier 3 assurance.
get_telemetry_statsGet aggregate pipeline telemetry statistics including phase timings, success rates, and throughput metrics across all jobs.
get_active_telemetryList currently active jobs with their current pipeline phase and telemetry status.
get_job_telemetryGet full event timeline for a specific job — all pipeline phase transitions with timings and metadata.
get_telemetry_logsQuery structured logs with filters for level, source, jobId, kernelId, time range, and full-text search.
emit_telemetryManually emit a telemetry event for a job pipeline phase. Used by kernels and agents to report execution progress.
submit_for_human_verificationSubmit an evidence bundle for human verification. Selects a panel of verifiers from the network and returns assigned verifier IDs.
get_verification_assignmentsGet pending verification assignments for a verifier node. Returns requests this verifier has been assigned but not yet responded to.
get_verification_statusGet verification status for a request. Returns vote tally, consensus state, dispute list, and pending response count.
respond_to_verificationSubmit a verifier's verdict for a verification request. Returns consensus state after this vote.
dispute_verificationFile a dispute against the consensus verdict. Only verifiers who submitted a response can dispute. Requires additional evidence CID.
register_capability_ipRegister a capability as a Story Protocol IP Asset. Returns an ipId and NFT token. The designer earns royalties whenever this capability is used.
register_job_evidence_ipRegister job evidence as a derivative IP asset on Story Protocol, linking it to the parent capability's IP. Operators earn royalties from derivative use.
distribute_royaltiesSet revenue splits for an IP asset among stakeholders (designer, operator, verifier, assembler, curator). Splits must sum to 100.
The first 60 of 256, as the server listed them.
How to connect
Add the server to your MCP client config. Claude Desktop, Cursor and most other clients read this shape:
{
"mcpServers": {
"physical-capability-cloud": {
"type": "http",
"url": "https://capability.network/mcp"
}
}
}Add it
Claude Code, in a terminal
claude mcp add --transport http capability https://capability.network/mcpclaude.ai, ChatGPT and other clients: add https://capability.network/mcp as a custom connector in the settings.
OpenAPI spec
733 paths, read 2026-10-26. Open the spec ↗
Each line is something a program found when it went looking: the registry entry, the server's own answer, a spec that parses. A missing line means we did not find it, not that it does not exist.
Cookies
Before we count this visit
Analytics and ad measurement tell us which pages and which of our ads do their job: Google Analytics and Google Ads, the Meta pixel, Yandex Metrica, and one first-party id of our own. Nothing is stored until you answer, and declining changes nothing about the site.