Skip to main content
Glama

Taifoon coordination layer

Grid compute and resources

taifoon_grid

Buys and sells measured compute and RPC on the Taifoon Grid: AI, GPU and TEE calls, hiring resources, joining with your own, earnings. For agent work use taifoon_find and taifoon_hire. Actions: status — what the Grid accepts and pays now, per resource kind; ai — one AI call on a node measured for that model; gpu — one inference through the metered GPU gate, with your key; tee — compile rules into typed questions inside an enclave; hire — usable providers of a kind with endpoints and a quote; join — bring a resource and earn GRID from measured probes; builders — join the Early Builder Program whitelist: get the message, sign it, send it; pool — where the Grid is short and what each kind pays; prices — every operation priced in and out in GRID; economics — what the Grid earns, per kind and attribution; settlement — a wallet's Grid ledger and what was minted; avenues — the outlets where Taifoon resells community capacity; shards — how the MMR splits across collectors; node — the docker command that runs taifoon-node for a wallet; bench — the one-line benchmark for a machine and its earnings.

Tool call template

This tool declares no required parameters. Add optional parameters from the input schema as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {},
    "name": "taifoon_grid"
  }
}

Input Schema

NameRequiredDescriptionDefault
idNostatus/release: the enclave id (enc_…) (actions: tee, avenues)
meNoyour libp2p peer id (base58) — returns your shard, config and command (actions: shards)
dayNoYYYYMMDD (UTC); default today (actions: economics)
refNoOptional: the 0x wallet that referred this resource. It earns a share of what the resource accrues (measured work only). Never the owner itself. (actions: join, node)
bodyNobody (actions: gpu)
gpusNonumber of cards (default 1) (actions: bench)
kindNokind (actions: hire, join, builders, node)
laneNolane (actions: tee)
nameNo2 to 64 characters (actions: builders)
nodeNodid:taifoon:z… of one node (actions: economics)
noteNooptional, ≤200 chars, stored with the intent (actions: hire)
poolNoLeave it out: the pool is Taifoon's side and is derived on the server (the Taifoon founding pool, fee 0). Kept only for nodes from before 2026-10-11 that still send one; honoured on first registration only. (actions: join)
sellNokind=gpu: true lists the node for sale at the BROKER's market price for its reference class (the rig never names a price; the cheapest other venue, input and output apart, sets it); the node is checked (battery + fingerprint) and owned by your key (actions: join)
agentNokind=agent: erc8004:<chain>:<id>, or the https URL of the agent card (actions: builders)
groupNogroup (actions: prices)
hirerNooptional 0x wallet; records the intent in the Grid ledger (actions: ai, hire)
hoursNohours (actions: tee, hire)
modelNothe model the node must serve (e.g. hypernova-60b, littlelamb) (actions: ai, gpu, join)
nonceNononce (actions: builders)
ownerNo0x wallet address that earns the GRID points (actions: join, economics, node, bench)
rulesNorules (actions: tee)
stateNostate (actions: prices)
unitsNohow many resources (default 1) (actions: hire)
actionNowhat to do (default status): status | ai | gpu | tee | hire | join | builders | pool | prices | economics | settlement | avenues | shards | node | benchstatus
expiryNoexpiry (actions: builders)
payoutNoOptional: grid (default, the full rate) or cash (USDC at the TSUL rate, 70 %) (actions: node)
promptNoyour prompt (one user message) (actions: ai)
walletNo0x wallet that joins (and later earns GRID and mints the pass) (actions: builders, settlement)
api_keyNoapi key (actions: gpu, tee, hire, join)
composeNoprovision/quote: an allowlisted compose id (taifoon-tee-seller) or its file’s sha256; absent = the first on the allowlist (actions: tee)
vram_gbNoVRAM of one card in GB (0 or omitted = no GPU) (actions: bench)
chain_idNosettlement chain (default 8453, Base — where the assurance contract is live) (actions: ai, join)
endpointNoendpoint (actions: gpu, join, builders, node)
providerNooptional: a specific measured endpoint from the providers list (actions: ai)
resourceNoonly avenues for this resource (actions: avenues)
referenceNokind=gpu: the market class your price is checked against (default llama-3.1-8b) (actions: join)
signatureNostep 2: the personal_sign of the message step 1 returned (actions: builders)
collectorsNoplan for this many collectors instead of the live swarm (actions: shards)
replicationNocollectors per shard (default 2) (actions: shards)

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed10 schema fields changed
    • changedInput schema / properties / action / description
      Previous value: -"what to do (default status): status | ai | gpu | tee | hire | join | pool | prices | economics | settlement | avenues | shards | node | bench"New value: +"what to do (default status): status | ai | gpu | tee | hire | join | builders | pool | prices | economics | settlement | avenues | shards | node | bench"
    • changedInput schema / properties / action / enum
      Previous value: -[
      -  "status",
      -  "ai",
      -  "gpu",
      -  "tee",
      -  "hire",
      -  "join",
      -  "pool",
      -  "prices",
      -  "economics",
      -  "settlement",
      -  "avenues",
      -  "shards",
      -  "node",
      -  "bench"
      -]New value: +[
      +  "status",
      +  "ai",
      +  "gpu",
      +  "tee",
      +  "hire",
      +  "join",
      +  "builders",
      +  "pool",
      +  "prices",
      +  "economics",
      +  "settlement",
      +  "avenues",
      +  "shards",
      +  "node",
      +  "bench"
      +]
    • addedInput schema / properties / agent
      Added value: +{
      +  "description": "kind=agent: erc8004:<chain>:<id>, or the https URL of the agent card (actions: builders)",
      +  "type": "string"
      +}
    • changedInput schema / properties / endpoint / description
      Previous value: -"endpoint (actions: gpu, join, node)"New value: +"endpoint (actions: gpu, join, builders, node)"
    • addedInput schema / properties / expiry
      Added value: +{
      +  "description": "expiry (actions: builders)",
      +  "type": "integer"
      +}
    • changedInput schema / properties / kind / description
      Previous value: -"kind (actions: hire, join, node)"New value: +"kind (actions: hire, join, builders, node)"
    • addedInput schema / properties / name
      Added value: +{
      +  "description": "2 to 64 characters (actions: builders)",
      +  "type": "string"
      +}
    • addedInput schema / properties / nonce
      Added value: +{
      +  "description": "nonce (actions: builders)",
      +  "type": "string"
      +}
    • addedInput schema / properties / signature
      Added value: +{
      +  "description": "step 2: the personal_sign of the message step 1 returned (actions: builders)",
      +  "type": "string"
      +}
    • changedInput schema / properties / wallet / description
      Previous value: -"optional 0x wallet (actions: settlement)"New value: +"0x wallet that joins (and later earns GRID and mints the pass) (actions: builders, settlement)"
  2. Changed1 schema field changed
    • changedInput schema / properties / endpoint / enum
      Previous value: -[
      -  "chat",
      -  "generate"
      -]New value: +[
      +  "chat",
      +  "generate",
      +  "route"
      +]
  3. Changed3 schema fields changed
    • removedInput schema / properties / price_in_usd_per_mtok
      Removed value: -{
      -  "description": "kind=gpu: your input price (default: the output price) (actions: join)",
      -  "type": "number"
      -}
    • removedInput schema / properties / price_out_usd_per_mtok
      Removed value: -{
      -  "description": "kind=gpu: YOUR price in USD per 1M output tokens; buyers pay it plus the reseller cut (actions: join)",
      -  "type": "number"
      -}
    • addedInput schema / properties / sell
      Added value: +{
      +  "description": "kind=gpu: true lists the node for sale at the BROKER's market price for its reference class (the rig never names a price; the cheapest other venue, input and output apart, sets it); the node is checked (battery + fingerprint) and owned by your key (actions: join)",
      +  "type": "boolean"
      +}
  4. Added

TDQS

B3.3/5.0
Behavior3/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare readOnlyHint=false, destructiveHint=false, openWorldHint=true and non-idempotence, so the safety baseline is supplied. The description adds that this is a commercial surface ('buys and sells', 'earnings', 'pays'), which usefully signals that several actions spend or commit funds. However it never says which actions are read-only versus financially committing, what authorization (api_key) each requires, or that join/sell register a node persistently — gaps that matter for a mutation-capable tool.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness3/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with purpose and the sibling hand-off, which is good structure. The action catalogue, however, is delivered as one dense semicolon-chained block of 15 items with no grouping, making it harder to scan than the information warrants. Each individual gloss is short, so there is little pure waste, but readability suffers.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness3/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

With 39 parameters, 100% schema coverage and no output schema, the description is obligated mainly to explain action selection and consequences, and it only partly does so. It omits return-value shape for the many read actions (status, pool, economics, settlement) and omits prerequisites such as which actions need api_key or a wallet. For a tool of this complexity it is adequate but leaves real gaps.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 100% and every parameter already carries a per-action applicability tag, so the schema does the heavy lifting and 3 is the baseline. The description's action glosses add context about what an action does (e.g. 'gpu — one inference through the metered GPU gate, with your key'), but not additional syntax or semantics for the parameters themselves. It does not compensate further, nor does it need to.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose4/5

Does the description clearly state what the tool does and how it differs from similar tools?

Opens with a concrete verb-plus-resource statement ('Buys and sells measured compute and RPC on the Taifoon Grid') and enumerates each of the 15 actions with a short gloss, so an agent can tell what the tool covers. It also names sibling alternatives ('For agent work use taifoon_find and taifoon_hire'), which separates it from the sibling set. It falls short of 5 because 15 actions are packed into one tool with uneven glosses — several (e.g. 'bench — the one-line benchmark for a machine and its earnings') remain too terse to be unambiguous.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines3/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

There is one explicit routing rule: agent work should go to taifoon_find and taifoon_hire, which is genuine when-not guidance. Beyond that, the agent is given no help choosing among closely related actions (status vs pool vs prices vs economics; join vs builders vs node), which is the main selection hazard for a 15-action tool. Usage is implied by the action list rather than stated as conditions.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.

Resources