API Reference
Complete reference for the ClawFeel relay API. Public endpoints require no authentication. Enterprise endpoints require a Bearer token.
https://api.clawfeel.ai
A Claw node reports its current Feel score, sensor data, and entropy hash to the relay. This is the primary ingestion endpoint for all network participants.
| Name | Type | Description |
|---|---|---|
| X-Claw-Id | string | Claw node identifier (e.g. Claw-a1b2c3d4). Optional — can also be sent in body. |
| Name | Type | Description |
|---|---|---|
| clawId | string | Node identifier. Required if not in header. |
| feel | number | Feel score (0–100). Required |
| era | string | Current era: Chaos, Transition, or Eternal. |
| hash | string | SHA-256 entropy hash (hex). |
| prevHash | string | Previous hash for chain verification. |
| seq | number | Sequence number (monotonically increasing). |
| authenticity | number | Sensor authenticity score (0–7). |
| entropyQuality | number | Entropy quality percentage (0–100). |
| sensorFlags | string | 7-bit string, each bit = sensor active (1111111). |
| alias | string | Human-readable node name. |
| type | string | node (hardware) or browser (light client). |
| sensors | object | Raw sensor readings (cpu, memory, etc.). |
| zkProof | object | Optional zero-knowledge proof for reputation boost. |
Returns a snapshot of the entire network: aggregate statistics, the computed network random number, latest beacon, and all currently online nodes.
Server-Sent Events (SSE) endpoint. Pushes network state updates every 2 seconds. Includes node data, aggregate stats, and beacon rounds.
Returns the DHT peer list for new nodes joining the network. Includes relay version, DHT port, and a list of active peers with their reputation scores.
Returns the most recently sealed beacon round. Beacons are sealed every 10 seconds by XOR-aggregating contributing node hashes, then SHA-256 hashing the result.
Retrieve a specific beacon round by its number. Returns 404 if the round has not been sealed yet or has been pruned.
| Name | Type | Description |
|---|---|---|
| round | integer | Beacon round number. Required |
Query a range of beacon rounds. Returns up to 100 rounds per request. Useful for historical analysis and verification.
| Name | Type | Description |
|---|---|---|
| from | integer | Start round (inclusive). Default: 1 |
| to | integer | End round (inclusive). Default: 999999 |
Search recent Feel report transactions. Supports filtering by hash prefix, clawId, or alias. Returns newest first.
| Name | Type | Description |
|---|---|---|
| search | string | Filter by hash prefix, clawId, or alias. |
| limit | integer | Max results (1–200). Default: 50 |
Look up a specific transaction by its hash prefix. The prefix must be at least 8 hex characters.
| Name | Type | Description |
|---|---|---|
| hash | string | Hash prefix (min 8 chars). Required |
Look up a specific node by clawId or alias. Supports partial matching. Returns full node state including online status and history.
| Name | Type | Description |
|---|---|---|
| id | string | ClawId or alias (supports partial match). Required |
Returns high-level network statistics: total online nodes, transaction count, and the latest beacon round number.
All endpoints below require a Bearer token. Get a free API key (100 req/hour) by registering:
Generate a single cryptographically secure random value derived from the latest beacon. Supports configurable bit length and output format. Uses HKDF-like entropy expansion from the beacon hash.
| Name | Type | Description |
|---|---|---|
| bits | integer | Entropy bits (8–4096). Default: 256 |
| format | string | Output format: hex, base64, decimal, bytes. Default: hex |
| Name | Type | Description |
|---|---|---|
| Authorization | string | Bearer <api_key> Required |
Generate multiple independent random values in a single request. Each value is derived from the same beacon round but with a unique index, ensuring statistical independence.
| Name | Type | Description |
|---|---|---|
| count | integer | Number of values (1–100). Default: 10 |
| bits | integer | Entropy bits per value (8–4096). Default: 256 |
| format | string | Output format: hex, base64, decimal, bytes. Default: hex |
| Name | Type | Description |
|---|---|---|
| Authorization | string | Bearer <api_key> Required |
Generate a uniformly distributed random integer within a specified range [min, max] (inclusive). Uses rejection sampling to eliminate modular bias.
| Name | Type | Description |
|---|---|---|
| min | integer | Minimum value (inclusive). Default: 1 |
| max | integer | Maximum value (inclusive). Default: 100. Must be > min. |
| Name | Type | Description |
|---|---|---|
| Authorization | string | Bearer <api_key> Required |
Verify a specific beacon round by recomputing the XOR accumulator and beacon hash from the original contributors. Returns whether all cryptographic checks pass.
| Name | Type | Description |
|---|---|---|
| round | integer | Beacon round number to verify. Required |
| Name | Type | Description |
|---|---|---|
| Authorization | string | Bearer <api_key> Required |
Check your API key's usage statistics: tier, rate limits, remaining requests in the current window, and total historical usage.
| Name | Type | Description |
|---|---|---|
| Authorization | string | Bearer <api_key> Required |
Let AI Agents make verifiable random decisions using hardware entropy. Works with LangChain, OpenAI function calling, AutoGPT, and any AI framework.
All methods return a result with a cryptographic proof object for audit trails.
| Method | Description | Example |
|---|---|---|
| decide(options) | Pick one randomly from array | agent.decide(['left', 'right']) |
| weighted(options, weights) | Weighted random selection | agent.weighted(['common', 'rare'], [80, 20]) |
| shuffle(array) | Cryptographic Fisher-Yates shuffle | agent.shuffle([1,2,3,4,5]) |
| sample(array, n) | Pick n unique items | agent.sample(players, 5) |
| probability(p) | True with probability p | agent.probability(0.3) |
| gaussian(mean, std) | Normal distribution | agent.gaussian(100, 15) |
Use ClawFeel as a native tool in LangChain, OpenAI function calling, or any framework that supports tool definitions.
Make agent decisions directly from the command line.
Every decision includes a cryptographic proof object for full auditability.