@atumlabs/payment-gateway-client 2.1.0 → 3.0.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,14 +1,17 @@
1
1
  # Payment Gateway Client
2
2
 
3
3
  TypeScript client for Atum Payment Gateway API with CLI support.
4
+ Full documentation: [docs.atum.xyz](https://docs.atum.xyz).
4
5
 
5
6
  > **Proprietary software - not open source.**
6
7
  >
7
8
  > Copyright (c) 2026 Atum Labs, Inc. All rights reserved. Access to and use of `@atumlabs/payment-gateway-client` are limited to entities expressly authorized by Atum and are governed by the Atum Payment Gateway Client Private Repository License Agreement in the LICENSE file included with this package. Do not redistribute, publish, mirror, sublicense, or provide this package to any unauthorized person. By accessing, installing, copying, or using the package, you agree to the license terms.
8
9
 
9
- > NOTE: Payments from EVM chains require prior ERC-20 approval on the source token for the Permit2 contract (0x000000000022D473030F116dDEE9F6B43aC78BA3 on virtually all EVM chains).
10
- >
11
- > (And the same applies for Tron/TRC-20, with a Permit2 address of TJhMXTHQHeQyMD7TcKQFqAePNgG4b31H9m on mainnet or TPUqJPASUn1zLvbLBgRZ5pBYrx7WSe5ahp on Shasta).
10
+ > NOTE: Before your first payment from an EVM or Tron chain, the source token must be approved
11
+ > for the Permit2 contract. Without it a payment is built, signed and accepted by the gateway,
12
+ > and then reverts on-chain at settlement. It is a one-off transaction per wallet and token —
13
+ > see [Approving the source token](#approving-the-source-token) for `ensureSourceApproval`,
14
+ > which does it for you.
12
15
 
13
16
  ## Installation
14
17
 
@@ -30,13 +33,40 @@ npm install -g @atumlabs/payment-gateway-client
30
33
  A project-local install also provides the commands, under `node_modules/.bin`, so
31
34
  `npx send-payment ...` works without a global install.
32
35
 
36
+ ## Supported corridors
37
+
38
+ A payment moves value from a source asset to a destination asset. The two can be on
39
+ different chains, and in different chain families.
40
+
41
+ Assets are named with [CAIP-19](https://chainagnostic.org/CAIPs/caip-19) identifiers:
42
+
43
+ | Chain family | Network | Example asset identifier |
44
+ | --- | --- | --- |
45
+ | EVM | Base Sepolia | `eip155:84532/erc20:0x036CbD53842c5426634e7929541eC2318f3dCF7e` |
46
+ | Solana | Devnet | `solana:EtWTRABZaYq6iMfeYKouRu166VU2xqa1/token:4zMMC9srt5Ri5X14GAgXhaHii3GnPAEERYPJgZJDncDU` |
47
+ | Tron | Shasta | `tron:shasta/trc20:TG3XXyExBkPp9nzdajDZsozEu4BkaSJozs` |
48
+
49
+ Every row above is a testnet, so nothing here settles real value if it is copied by mistake.
50
+ The mainnet identifier for the same asset is a different string on every chain — take it from
51
+ [supported assets](https://docs.atum.xyz/get-started/reference/supported-assets), never by
52
+ editing one of these.
53
+
54
+ For the identifier of each supported token and network, see
55
+ [supported assets](https://docs.atum.xyz/get-started/reference/supported-assets) and
56
+ [supported networks](https://docs.atum.xyz/get-started/reference/supported-networks).
57
+
58
+ Copy identifiers exactly: base58 values, such as Solana chain references, token mints
59
+ and account addresses, are case-sensitive. The escrow and verifier addresses for a
60
+ chain are fetched by `preparePaymentRequest`, so you never supply them yourself.
61
+
33
62
  ## CLI Usage
34
63
 
35
- The package provides four commands: `send-payment`, `request-quote`,
64
+ The package provides five commands: `send-payment`, `approve-permit2`, `request-quote`,
36
65
  `quote-status`, and `payment-status`. After a global install:
37
66
 
38
67
  ```bash
39
68
  send-payment [options] <sender> <amount> <sourceAsset> <receiver> <destinationAsset>
69
+ send-payment --submit <file> [options]
40
70
  ```
41
71
 
42
72
  ### Arguments
@@ -51,19 +81,59 @@ send-payment [options] <sender> <amount> <sourceAsset> <receiver> <destinationAs
51
81
 
52
82
  - `-V, --version` - Output the version number
53
83
  - `--gateway <url>` - Payment gateway URL (default: testnet)
54
- - `--request-id <id>` - Idempotency key; seeds the on-chain deposit nonce. Reuse the same id to retry a payment safely. Omitted: a one-shot id is generated and printed (see [Idempotency and retries](#idempotency-and-retries))
84
+ - `--request-id <id>` - Idempotency key; seeds the on-chain deposit nonce. Re-run with the same id to re-attempt one payment. Omitted, an id is generated and printed, which makes every bare re-run a separate payment (see [Idempotency and retries](#idempotency-and-retries))
55
85
  - `--request-id-prefix <prefix>` - Prefix for the auto-generated request id (default: `pmt-gw-cli`)
86
+ - `--prepare-only` - Build and sign the request, print it to stdout, and stop without submitting. For inspecting an authorization, or for review before paying; send it later with `--submit` (see [Preparing a payment for review](#preparing-a-payment-for-review)). Cannot be combined with `--wait`
87
+ - `--submit <file>` - Send a request built earlier by `--prepare-only`, instead of building one. Signs nothing and needs no private key; `-` reads stdin. Takes no corridor arguments
88
+ - `--wait` - After submitting, keep checking until the payment settles or fails instead of reporting it as still pending
89
+ - `--wait-seconds <n>` - How long `--wait` keeps checking (default: `60`). A convenience bound only; giving up does not affect the payment
56
90
  - `-h, --help` - Display help
57
91
 
58
- ### CAIP-19 Asset Format Examples
92
+ ### Exit codes
93
+
94
+ `send-payment` reports the payment's ending, so `send-payment ... && <next step>`
95
+ runs the next step only for a settled payment:
96
+
97
+ | Code | Meaning |
98
+ | ---- | --------------------------------------------------------------------------- |
99
+ | `0` | `completed` — settled |
100
+ | `2` | `pending` — accepted, no result yet |
101
+ | `3` | `failed` / `cancelled` — terminal; a fresh attempt needs a new `--request-id` |
102
+ | `1` | no outcome obtained (rejected, gateway unreachable, bad arguments) |
103
+
104
+ On `2`, follow the payment with **`payment-status <payment_id>`** — the id is in the
105
+ output — or use `--wait` to have `send-payment` do it for you. Re-running
106
+ `send-payment` with the same `--request-id` also returns the current state, but it
107
+ rebuilds and re-sends the payment to do so, which is more work than a status
108
+ lookup.
109
+
110
+ `1` is the case that reverses that order. It means the outcome is **unknown**, not
111
+ that nothing happened: a request that timed out may still have been accepted, and
112
+ you may not have a `payment_id` to look up. Re-run with the same `--request-id`,
113
+ which resolves to the payment if one was created and creates it if not. The id is
114
+ printed on failure for exactly this.
59
115
 
60
- - **EVM ERC20**: `eip155:42161/erc20:0xaf88d065e77c8cC2239327C5EDb3A432268e5831`
61
- - **Tron TRC20**: `tron:shasta/trc20:TG3XXyExBkPp9nzdajDZsozEu4BkaSJozs`
116
+ `payment-status` is a lookup, so its exit code reports whether the lookup worked
117
+ (`0`) or not (`1`), never the payment's outcome — read `status` for that.
118
+
119
+ **The table describes a submitted payment, so it does not apply to `--prepare-only`.**
120
+ There, `0` means the request was built and signed — nothing was submitted and no
121
+ payment exists. Do not chain a payment-conditional step off a `--prepare-only` run.
122
+
123
+ ### Asset format
124
+
125
+ `sourceAsset` and `destinationAsset` are CAIP-19 identifiers — see
126
+ [Supported corridors](#supported-corridors).
62
127
 
63
128
  ### Signing provider
64
129
 
65
130
  Choose how the sender authorization is signed with `--provider` (default: `raw`).
66
131
 
132
+ Whichever provider you use, pass `pinnedAddress` to `createSenderSigner` with the payment's
133
+ source account. The SDK checks that the signing key resolves to that address before anything
134
+ is signed, so the wrong key fails immediately instead of producing a signature the gateway
135
+ rejects. This works on every chain and on both providers.
136
+
67
137
  #### Private key (`--provider raw`, default)
68
138
 
69
139
  Provide your private key via environment variable:
@@ -97,7 +167,7 @@ does not pull in Turnkey or `@solana/web3.js`.
97
167
  The command-line tools also read an optional `.env` file from the directory you
98
168
  run them in, so credentials and settings do not have to be retyped each time:
99
169
 
100
- ```
170
+ ```bash
101
171
  PRIVATE_KEY=0x...
102
172
  GATEWAY_URL=https://payment-gw.production-testnet.atum.xyz
103
173
  SOLANA_RPC_URL=
@@ -120,6 +190,10 @@ environment.
120
190
  - **PaymentRequest JSON** → `stderr` (for debugging/logging)
121
191
  - **PaymentResponse JSON** → `stdout` (for parsing/automation)
122
192
 
193
+ Under `--prepare-only` there is no response, and the signed request is the result, so
194
+ it goes to `stdout` instead and the `stderr` copy is dropped. That keeps
195
+ `send-payment --prepare-only ... | jq` working.
196
+
123
197
  ### Example
124
198
 
125
199
  ```bash
@@ -131,6 +205,103 @@ PRIVATE_KEY=0xabc... send-payment \
131
205
  tron:shasta/trc20:TG3XXyExBkPp9nzdajDZsozEu4BkaSJozs
132
206
  ```
133
207
 
208
+ ### Preparing a payment for review
209
+
210
+ `send-payment` builds, signs and submits in one step. To have the payment approved
211
+ before it is sent, split it in two: `--prepare-only` stops after signing and prints
212
+ the request instead, so you can see exactly what you are about to authorize — the
213
+ document is not a description of the payment, it *is* the payment. `--submit` sends
214
+ one that was prepared earlier.
215
+
216
+ ```bash
217
+ # 1. build and sign, submit nothing
218
+ send-payment --prepare-only --quote-deadline-seconds 30 \
219
+ 0xF0D0... 100000 \
220
+ eip155:42161/erc20:0xaf88d065e77c8cC2239327C5EDb3A432268e5831 \
221
+ 0xMerchant... \
222
+ eip155:42220/erc20:0xcebA9300f2b948710d2653dD7B07f33A8B32118C \
223
+ > request.json
224
+
225
+ # 2. inspect it, have it reviewed
226
+
227
+ # 3. send it
228
+ send-payment --submit request.json --wait
229
+ ```
230
+
231
+ Or piped, with no file on disk:
232
+
233
+ ```bash
234
+ send-payment --prepare-only <args...> | send-payment --submit -
235
+ ```
236
+
237
+ Four things to know:
238
+
239
+ - **`--prepare-only` still contacts the gateway.** The corridor addresses come from
240
+ it, so this is not an offline signing mode.
241
+ - **Widen the quote window if you intend to submit later — but only by tens of seconds.**
242
+ A prepared document is submittable only while its quote window is open, and the default
243
+ is 10 seconds. That window is also the auction: the gateway waits it out before choosing
244
+ a settler, and settler quotes expire on their own, so an over-long window ends in
245
+ `QUOTES_EXPIRED`. That failure is terminal, and a fresh attempt needs a new request id.
246
+ Give the reviewer enough time and no more.
247
+ - **`--submit` signs nothing and needs no private key** — the authorization is already
248
+ in the file. It also cannot repair one: a request whose signature does not match its
249
+ contents is rejected, and the fix is to prepare a new one.
250
+ - **Treat the prepared file like a signed cheque.** It is a signed authorization, not a
251
+ draft: anyone who has it can send it, because `--submit` needs no key and the signature
252
+ is already inside. Keep it where you would keep a credential, and delete it once the
253
+ payment is sent or the request is abandoned.
254
+
255
+ `--submit` takes no corridor arguments, since the request already carries them, and it
256
+ rejects the options that only shape a request being built (`--request-id`,
257
+ `--quote-deadline-seconds`, `--provider` and the rest) rather than accepting and ignoring
258
+ them. It reports with the same exit codes and streams as an ordinary run.
259
+
260
+ Re-preparing with the same `--request-id` targets the **same payment** — the deposit nonce
261
+ is derived from the id, so it cannot pay twice — but it does not reproduce the same bytes.
262
+ The deadlines are stamped from the clock on each run and the signature covers them, so two
263
+ prepared documents for one payment always differ in `quote_deadline`,
264
+ `fulfillment_deadline` and the signature.
265
+
266
+ ### Approving the source token from the CLI
267
+
268
+ `approve-permit2` grants the one-off approval described in
269
+ [Approving the source token](#approving-the-source-token). Run it once per wallet and token,
270
+ before the first `send-payment` from that token:
271
+
272
+ ```bash
273
+ approve-permit2 <sender> <sourceAsset> --rpc-url <url>
274
+ ```
275
+
276
+ ```bash
277
+ # is one needed at all? reads the allowance, sends nothing, costs no gas
278
+ approve-permit2 0xF0D0... eip155:421614/erc20:0xaf88... --rpc-url https://... --check
279
+
280
+ # grant it
281
+ approve-permit2 0xF0D0... eip155:421614/erc20:0xaf88... --rpc-url https://...
282
+ ```
283
+
284
+ This is the only command in the package that broadcasts a transaction and spends gas.
285
+
286
+ - `--rpc-url <url>` - RPC endpoint for the source chain. Also read from `RPC_URL`.
287
+ - `--check` - report whether an approval is needed, then exit. Needs no private key.
288
+ - `--amount <units>` - approve only this many smallest units instead of an unlimited approval.
289
+ - `--spender <address>` - approve a contract other than the canonical Permit2.
290
+
291
+ The private key comes from `PRIVATE_KEY` or an interactive prompt, never a flag — a flag would
292
+ put it in your shell history and in the process list. It is only requested once the allowance has
293
+ been read and an approval is genuinely due, so an already-approved wallet never has to produce
294
+ one, and it is checked against `<sender>` before anything is sent. The result JSON goes to stdout.
295
+
296
+ EVM and Tron. Solana needs no approval at all, and the command says so rather than doing nothing.
297
+
298
+ Tron additionally needs the `tronweb` package, which is an **optional** peer dependency so that
299
+ EVM-only installs do not carry it. Install it alongside the CLI when you pay from Tron:
300
+
301
+ ```bash
302
+ npm install -g tronweb
303
+ ```
304
+
134
305
  ### Removing the command-line tools
135
306
 
136
307
  ```bash
@@ -147,7 +318,6 @@ import {
147
318
  createSenderSigner,
148
319
  signPaymentRequest,
149
320
  getErrorResponse,
150
- type PaymentResponse,
151
321
  ApiError,
152
322
  } from '@atumlabs/payment-gateway-client';
153
323
 
@@ -167,6 +337,27 @@ environment, in order of precedence:
167
337
 
168
338
  Mainnet access is granted by Atum; use the mainnet gateway URL provided to you.
169
339
 
340
+ ### Timeouts
341
+
342
+ Requests are bounded so an unresponsive gateway fails rather than hanging. There
343
+ are two budgets, because some endpoints are held open by the gateway on purpose
344
+ while it collects quotes or settles a payment:
345
+
346
+ | Budget | Default | Applies to |
347
+ | --- | --- | --- |
348
+ | `timeoutMs` | 30s | requests the gateway answers as soon as it can |
349
+ | `syncWaitTimeoutMs` | 120s | `POST /v1/payments`, `POST /v1/payments/awards`, and any `wait: true` call |
350
+
351
+ ```typescript
352
+ const client = new PaymentGatewayClient({ timeoutMs: 10_000, syncWaitTimeoutMs: 60_000 });
353
+ ```
354
+
355
+ Exceeding either raises a `GatewayTimeoutError` (see Error Handling). Raise
356
+ `syncWaitTimeoutMs` if your gateway is configured with a larger
357
+ `payment_sync_wait_seconds` than the 30s default, since it must clear that cap.
358
+ For a single request, an interceptor that sets `timeout` takes precedence over
359
+ both.
360
+
170
361
  ### Example: Tron to Celo Sepolia Payment
171
362
 
172
363
  ```typescript
@@ -207,13 +398,36 @@ async function sendPayment() {
207
398
  // Submit the payment
208
399
  const response = await client.payments.submitPayment({
209
400
  requestBody: paymentRequest
210
- }) as PaymentResponse;
401
+ });
211
402
 
212
- console.log('Payment successful!');
403
+ // A 2xx means the payment was accepted, NOT that it settled. Branch on `status`:
404
+ // cross-chain settlement can outrun the gateway's synchronous window, so an accepted
405
+ // payment may still have no result yet. See "Idempotency and retries" below.
213
406
  console.log('Payment ID:', response.payment_id);
214
- console.log('Source TX:', response.fulfillment_confirmation?.source_tx_hash);
215
- console.log('Destination TX:', response.fulfillment_confirmation?.destination_tx_hash);
216
-
407
+ if (response.idempotent_replay) {
408
+ // This request_id had been used before, so this is the payment your earlier attempt
409
+ // created. Nothing additional was charged.
410
+ console.log('Handed back an existing payment');
411
+ }
412
+
413
+ switch (response.status) {
414
+ case 'completed':
415
+ console.log('Settled.');
416
+ console.log('Source TX:', response.fulfillment_confirmation?.source_tx_hash);
417
+ console.log('Destination TX:', response.fulfillment_confirmation?.destination_tx_hash);
418
+ break;
419
+ case 'failed':
420
+ case 'cancelled':
421
+ // Terminal, and this request_id stays bound to it. A fresh attempt needs a NEW id.
422
+ console.error('Failed:', response.error?.code, response.error?.message);
423
+ break;
424
+ default:
425
+ // Still settling. Do not release goods, and do not pay again under a new id.
426
+ // Follow it with getPaymentStatus(payment_id); re-submitting this same request
427
+ // also returns the current state, but it re-sends the payment to do so.
428
+ console.log('Still settling — see "Tracking a payment" below.');
429
+ }
430
+
217
431
  return response;
218
432
  } catch (error) {
219
433
  const gwError = getErrorResponse(error);
@@ -227,16 +441,181 @@ async function sendPayment() {
227
441
  }
228
442
  ```
229
443
 
444
+ ### Tracking a payment
445
+
446
+ A 2xx from `submitPayment` means the payment was accepted, not that it settled. Read
447
+ `status` on the response, and follow the payment while it is still `pending`.
448
+
449
+ **Store your `request_id` before you call `submitPayment`, not after.** It is the only
450
+ handle that survives a crash between sending the request and receiving the response, and
451
+ it is what you need to find the payment again.
452
+
453
+ | Status | Meaning |
454
+ | --- | --- |
455
+ | `pending` | Accepted, not yet settled |
456
+ | `completed` | Delivered |
457
+ | `failed` | Did not settle |
458
+ | `cancelled` | Cancelled before completion |
459
+
460
+ `completed`, `failed` and `cancelled` are final. A payment that reaches one of them keeps
461
+ its `request_id` for good, so re-submitting that id returns the same finished payment
462
+ rather than trying again — a genuine new attempt needs a **new** `request_id`. Any status
463
+ this client does not recognise is treated as still in flight. `isTerminalStatus` is exported
464
+ so you can ask that question without keeping your own copy of the list: a value added to the
465
+ contract then cannot silently read as finished in your code.
466
+
467
+ `getPaymentStatus` returns the payment's state at that moment:
468
+
469
+ ```typescript
470
+ import { isTerminalStatus } from '@atumlabs/payment-gateway-client';
471
+
472
+ const snapshot = await client.payments.getPaymentStatus({ paymentId });
473
+
474
+ if (!isTerminalStatus(snapshot.status)) {
475
+ // Still in flight. Do not release goods.
476
+ }
477
+
478
+ if (snapshot.error) {
479
+ // Present when the payment failed. The chain, transaction hash and revert reason are
480
+ // under `blockchain_context`, not on the error itself.
481
+ console.error(
482
+ snapshot.error.code,
483
+ snapshot.error.message,
484
+ snapshot.error.blockchain_context?.tx_hash ?? '',
485
+ );
486
+ }
487
+ ```
488
+
489
+ #### Waiting for an outcome
490
+
491
+ `getPaymentStatus` answers once and does not wait. To follow a payment to its end, use
492
+ `waitForTerminalStatus`, which is the same helper behind the `send-payment --wait` flag:
493
+
494
+ ```typescript
495
+ import { waitForTerminalStatus } from '@atumlabs/payment-gateway-client';
496
+
497
+ const outcome = await waitForTerminalStatus({
498
+ fetchStatus: () => client.payments.getPaymentStatus({ paymentId }),
499
+ budgetMs: 60_000,
500
+ onError: (error) => console.warn('status check failed, still waiting:', error),
501
+ });
502
+
503
+ if (outcome.timedOut) {
504
+ // Still in flight when the budget ran out. Not a failure — check again later.
505
+ } else {
506
+ console.log(outcome.snapshot?.status);
507
+ }
508
+ ```
509
+
510
+ A failed status check does not end the wait: a network blip while a payment is settling is
511
+ the worst moment to stop looking, so failures are counted in `outcome.errorCount`, reported
512
+ through `onError`, and polling continues.
513
+
514
+ `budgetMs` defaults to `DEFAULT_WAIT_BUDGET_MS` (one minute) and is in **milliseconds**.
515
+
516
+ Checks are `DEFAULT_POLL_INTERVAL_MS` apart
517
+ (2 seconds) by default; pass `intervalMs` to widen it, and if you are following many
518
+ payments at once, stagger their start times so their polls do not align.
519
+
520
+ Pass a `signal` to stop a wait early, for a process draining on shutdown or a request that
521
+ has been cancelled. The outcome then reports `aborted`, and still carries the last status it
522
+ read. `timedOut` is true whenever the wait ended without a terminal status, so branching on
523
+ it stays correct either way; read `aborted` when you need to tell the two apart.
524
+
525
+ ```typescript
526
+ const controller = new AbortController();
527
+ process.once('SIGTERM', () => controller.abort());
528
+
529
+ const outcome = await waitForTerminalStatus({
530
+ fetchStatus: () => client.payments.getPaymentStatus({ paymentId }),
531
+ signal: controller.signal,
532
+ });
533
+
534
+ if (outcome.aborted) {
535
+ // Shutting down. The payment is unaffected; pick it up again on the next start.
536
+ }
537
+ ```
538
+
539
+ Running out of budget says nothing about the payment: `timedOut` means you stopped
540
+ watching, not that anything failed. Check again later rather than paying again — see
541
+ [Idempotency and retries](#idempotency-and-retries).
542
+
543
+ #### Finding a payment again
544
+
545
+ If a submit call threw, you may hold no `payment_id` — but the payment may still have been
546
+ created. `resolvePayment` turns the `request_id` you stored back into one. It throws an
547
+ `ApiError` with status `404` when the gateway has no payment recorded for that id:
548
+
549
+ ```typescript
550
+ import { isApiError } from '@atumlabs/payment-gateway-client';
551
+
552
+ try {
553
+ const { payment_id } = await client.payments.resolvePayment({ requestId: 'order-4711' });
554
+ const snapshot = await client.payments.getPaymentStatus({ paymentId: payment_id });
555
+ console.log(snapshot.status);
556
+ } catch (error) {
557
+ if (isApiError(error) && error.status === 404) {
558
+ // No payment recorded for that request_id yet.
559
+ } else {
560
+ throw error;
561
+ }
562
+ }
563
+ ```
564
+
565
+ A `404` is not proof that nothing happened. A submit that timed out may still be in flight,
566
+ and the gateway records the payment when it lands. **Never start a fresh payment under a new
567
+ `request_id` to recover** — that is a second payment, with real money on both. Re-submit the
568
+ **same** `request_id` instead: it returns the payment if one was created, and creates it if
569
+ not.
570
+
571
+ #### Step-by-step detail
572
+
573
+ `getPaymentTimeline` returns the payment's individual steps, each with its own status and,
574
+ for on-chain steps, a chain id and transaction hash. Use it to show progress, or to
575
+ investigate a payment that did not settle:
576
+
577
+ ```typescript
578
+ const timeline = await client.payments.getPaymentTimeline({ paymentId });
579
+
580
+ for (const step of timeline.steps) {
581
+ console.log(step.status, step.label, step.tx_hash ?? '');
582
+ }
583
+
584
+ if (timeline.error) {
585
+ // Why it failed, and where: for an on-chain failure this carries the chain id, the
586
+ // transaction hash and the revert reason.
587
+ console.error(timeline.error.code, timeline.error.message, timeline.error.revert_reason ?? '');
588
+ }
589
+ ```
590
+
591
+ The timeline reports `settling` while on-chain steps are running, for a payment that
592
+ `getPaymentStatus` still reports as `pending`.
593
+
594
+ ### Solana signing
595
+
596
+ Solana signs with Ed25519, which does not let a verifier recover the signer's key from
597
+ the signature, so the key travels with the message as `payload.delegate_signer`.
598
+ `signPaymentRequest` and the command-line tools set it for you.
599
+
600
+ The key you sign with must be the payment's source account, which is what `pinnedAddress`
601
+ checks — see [Signing provider](#signing-provider).
602
+
230
603
  ### Error Handling
231
604
 
232
605
  The gateway returns a typed, public error body — `ErrorResponse`
233
- (`{ code, message, request_id?, docs_url?, domain? }`). Use `getErrorResponse`
234
- to narrow a caught error to it; it returns `undefined` for anything that is not
235
- a gateway error with that shape. For transport-level detail (HTTP status, URL),
236
- the thrown value is still an `ApiError`.
606
+ (`{ code, message, request_id?, payment_id?, docs_url?, domain? }`). Use
607
+ `getErrorResponse` to narrow a caught error to it; it returns `undefined` for
608
+ anything that is not a gateway error with that shape. `payment_id` is present only
609
+ when the error is about a specific payment, so it is absent on the validation and
610
+ authentication failures that happen before a payment exists. For transport-level
611
+ detail (HTTP status, URL), the thrown value is still an `ApiError`.
612
+
613
+ Input validation happens before any request, so those failures are a plain `Error` carrying the
614
+ message and no `ErrorResponse` — `getErrorResponse` returns `undefined` for them. Read
615
+ `error.message`.
237
616
 
238
617
  ```typescript
239
- import { getErrorResponse, ApiError } from '@atumlabs/payment-gateway-client';
618
+ import { getErrorResponse, isGatewayTimeoutError, ApiError } from '@atumlabs/payment-gateway-client';
240
619
 
241
620
  try {
242
621
  await client.payments.submitPayment({ requestBody: paymentRequest });
@@ -247,6 +626,10 @@ try {
247
626
  console.error('Message:', gwError.message); // safe, human-readable text
248
627
  console.error('Request ID:', gwError.request_id); // quote this to support
249
628
  console.error('Docs:', gwError.docs_url);
629
+ } else if (isGatewayTimeoutError(error)) {
630
+ // The gateway never answered within the client's budget. There is no
631
+ // response and no status, so this is not an ApiError.
632
+ console.error(error.message, error.timeoutMs);
250
633
  } else if (error instanceof ApiError) {
251
634
  // Transport-level failure with no typed body (e.g. a proxy/gateway error).
252
635
  console.error('HTTP', error.status, error.statusText, error.url);
@@ -256,36 +639,217 @@ try {
256
639
  }
257
640
  ```
258
641
 
642
+ A thrown error does **not** mean no payment was created — a request that timed out
643
+ may still have been accepted. Treat it as unknown rather than as a failure:
644
+
645
+ - If you already hold the `payment_id` from an earlier attempt, read
646
+ `getPaymentStatus` — the cheapest way to find out what happened.
647
+ - If you do not, call `resolvePayment` with the same `request_id` to get the
648
+ `payment_id`, then read its status. A `404` there means the gateway has no payment
649
+ recorded for that id *yet* — see [Finding a payment again](#finding-a-payment-again).
650
+ - Re-submitting the same `request_id` also works and needs no lookup: it resolves to
651
+ the payment if one was created, and creates it if not. Never recover by starting a
652
+ payment under a **new** `request_id`: that is a second payment, not a retry.
653
+
654
+ One code is worth handling by name:
655
+
656
+ ```typescript
657
+ const gwError = getErrorResponse(error);
658
+ if (gwError?.code === 'IDEMPOTENCY_TERMS_MISMATCH') {
659
+ // 409 — this request_id already identifies a payment with different economics.
660
+ // Re-submitting cannot amend that payment; a genuinely different one needs a new
661
+ // request_id. See "Idempotency and retries".
662
+ }
663
+ ```
664
+
259
665
  ### Asset Identifiers
260
666
 
261
- Assets are identified using the [CAIP-19](https://chainagnostic.org/CAIPs/caip-19) format:
262
- - Format: `{chain_id}/{asset_namespace}:{asset_reference}`
263
- - Tron example: `tron:shasta/trc20:TG3XXyExBkPp9nzdajDZsozEu4BkaSJozs`
264
- - EVM example: `eip155:11142220/erc20:0x456a3D042C0DbD3db53D5489e98dFb038553B0d0`
667
+ Assets are identified using the [CAIP-19](https://chainagnostic.org/CAIPs/caip-19) format,
668
+ `{chain_id}/{asset_namespace}:{asset_reference}` — see
669
+ [Supported corridors](#supported-corridors) for an example per chain family.
670
+
671
+ Identifiers are validated before a request is built, so a malformed one is reported with the value
672
+ and what to correct rather than surfacing from a chain library. The rules: both namespaces are
673
+ canonicalized to lowercase, an `eip155` chain id is decimal, an EVM token address is 20-byte hex
674
+ and, when mixed-case, must carry a valid [EIP-55](https://eips.ethereum.org/EIPS/eip-55) checksum,
675
+ and a native asset (`slip44`) has no contract to deposit so it is refused. `parseAssetIdentifier`
676
+ and `assetChainId` are exported if you want the same parsing and canonicalization in your own code.
677
+
678
+ The checksum rule applies to the destination as well as the source, which is stricter than the
679
+ gateway, and it is typo protection rather than a settlement guard. For a destination address you
680
+ cannot edit, the all-lowercase form is the same twenty bytes and is accepted. Do not reach for that
681
+ on the source: settler configuration matches the source asset identifier byte for byte, so
682
+ lowercasing it moves the failure to settlement instead of removing it.
683
+
684
+ ## Approving the source token
685
+
686
+ On EVM and Tron the escrow moves your token through **Permit2**, so the token contract has to be
687
+ told that Permit2 may move it. That is a one-off transaction per `(wallet, token)` — every payment
688
+ afterwards is just an offline signature and costs nothing.
689
+
690
+ Skipping it produces the worst kind of failure: the request is built correctly, signed correctly
691
+ and accepted by the gateway, and then settlement reverts on-chain minutes later.
692
+
693
+ `ensureSourceApproval` reads the current allowance and sends an approval only if it falls short:
694
+
695
+ ```ts
696
+ import { ensureSourceApproval, needsSourceApproval } from '@atumlabs/payment-gateway-client';
697
+ import { Wallet, JsonRpcProvider } from 'ethers';
698
+
699
+ const wallet = new Wallet(process.env.PRIVATE_KEY!, new JsonRpcProvider(RPC_URL));
700
+
701
+ const result = await ensureSourceApproval({
702
+ network: 'eip155:11142220',
703
+ token: '0x456a3D042C0DbD3db53D5489e98dFb038553B0d0',
704
+ owner: wallet.address,
705
+ signer: wallet,
706
+ });
707
+
708
+ if (!result.alreadySufficient) {
709
+ console.log(`approved in ${result.txHash}`);
710
+ }
711
+ ```
712
+
713
+ You supply the signer, so the SDK never picks an RPC endpoint for you — everything else in this
714
+ package is offline, and this is the one call that broadcasts a transaction and spends gas.
715
+
716
+ It waits for the approval to be mined before returning, so `txHash` means confirmed rather than
717
+ merely sent. That wait is bounded — one minute by default, `confirmation.timeoutMs` to change it —
718
+ and a timeout is reported as *unconfirmed*, not failed, because the transaction may still land.
719
+ Pass `onSubmitted` to receive the hash the moment it is broadcast, so a slow approval can still be
720
+ looked up:
721
+
722
+ ```ts
723
+ await ensureSourceApproval({
724
+ network, token, owner, signer,
725
+ onSubmitted: (txHash) => console.log(`watching ${txHash}`),
726
+ confirmation: { timeoutMs: 120_000 },
727
+ });
728
+ ```
729
+
730
+ `isUnconfirmed(error)` tells the two apart. It matters because the responses are opposite: a
731
+ failed approval needs another one, an unconfirmed approval needs a look at the transaction and
732
+ nothing else. Sending a second one on top of the first only pays for an allowance you are already
733
+ getting.
734
+
735
+ ```ts
736
+ try {
737
+ await ensureSourceApproval({ network, token, owner, signer });
738
+ } catch (error) {
739
+ if (isUnconfirmed(error)) {
740
+ // Broadcast, not yet seen to confirm. Check the hash from onSubmitted; do not re-send.
741
+ } else {
742
+ throw error;
743
+ }
744
+ }
745
+ ```
746
+
747
+ On EVM, some tokens — mainnet USDT is the one you are most likely to meet — refuse to move an
748
+ allowance straight from one non-zero value to another. If your wallet already holds a partial
749
+ allowance on such a token, the approval is reset to zero first, automatically, and the result
750
+ carries a `resetTxHash` alongside the usual `txHash`. The check costs nothing and sends nothing,
751
+ and a wallet with no allowance skips it entirely, so an ordinary approval still takes exactly one
752
+ transaction.
753
+
754
+ Tron does not do this. An ordinary TRC-20 accepts the change in place, so guessing otherwise would
755
+ burn an extra transaction at Tron's fee limit; if a Tron token ever does refuse, the error tells
756
+ you to approve zero first and retry.
757
+
758
+ Paying from Tron takes a TronWeb instance in place of the ethers signer. The Permit2 address for
759
+ the network is resolved for you:
760
+
761
+ ```ts
762
+ await ensureSourceApproval({ network: 'tron:shasta', token, owner, tronWeb });
763
+ ```
764
+
765
+ Solana needs no approval at all — the signed deposit authorizes the transfer itself — so the call
766
+ reports `alreadySufficient` without touching the chain.
767
+
768
+ ### Checking without spending gas
769
+
770
+ `needsSourceApproval` takes the same arguments and answers the same question, read-only. Use it to
771
+ warn a payer that a transaction is coming before asking them to sign:
772
+
773
+ ```ts
774
+ if (await needsSourceApproval({ network: 'eip155:11142220', token, owner, signer })) {
775
+ // a one-off approval transaction is due first
776
+ }
777
+ ```
778
+
779
+ ### Bounded approvals
780
+
781
+ The approval sent is unlimited by default, which is why later payments need no further
782
+ transaction. Pass `approvalAmount` to bound the approval this sends instead.
783
+
784
+ On its own the amount is treated as its own requirement, so an allowance still holding the full
785
+ bound is left alone rather than re-approved — the right answer for a one-off approval. Across
786
+ repeated payments, pair it with `requiredAllowance`, as below: Permit2 **decrements** the
787
+ allowance on every payment, so a bound measured against itself stops covering itself the moment
788
+ the first charge lands, and every later call would send another approval. Either way a bounded
789
+ approval is consumed as it is spent and eventually has to be granted again:
790
+
791
+ ```ts
792
+ await ensureSourceApproval({
793
+ network: 'eip155:11142220',
794
+ token,
795
+ owner,
796
+ signer,
797
+ requiredAllowance: 1_000_000n, // what this payment needs
798
+ approvalAmount: 5_000_000n, // approve only this much, not unlimited
799
+ });
800
+ ```
801
+
802
+ These helpers raise an allowance to what a payment needs; they never lower one. A wallet already
803
+ holding more than `approvalAmount` is left exactly as it is and reports `alreadySufficient`
804
+ without sending anything. Reducing or revoking an allowance is a separate operation, and not one
805
+ these perform.
806
+
807
+ Permit2 contract addresses, if you would rather approve by hand:
808
+
809
+ | Chain | Address |
810
+ | --- | --- |
811
+ | EVM (virtually all chains) | `0x000000000022D473030F116dDEE9F6B43aC78BA3` |
812
+ | Tron mainnet | `TJhMXTHQHeQyMD7TcKQFqAePNgG4b31H9m` |
813
+ | Tron Shasta | `TPUqJPASUn1zLvbLBgRZ5pBYrx7WSe5ahp` |
265
814
 
266
815
  ## Idempotency and retries
267
816
 
268
- Every payment carries a `request_id`. Pass a stable value and reuse the exact
269
- same one on every retry (a timeout, a network blip, a restart) and you end up
270
- with one payment, not two — and you get the same answer each time.
271
-
272
- - Using this SDK, `request_id` is **required** on the payment path
273
- (`preparePaymentRequest`), and it is the only thing you manage: the SDK derives
274
- the on-chain deposit nonce deterministically from it, so retries line up on
275
- their own. Reuse the same `request_id` to retry safely.
276
- - On the CLI, `--request-id` is optional; when omitted a one-shot id is generated
277
- and printed, so you can re-run with `--request-id <id>` to retry the same
278
- payment.
279
- - A genuinely new payment needs a **new** `request_id`. Reusing one with different
280
- terms (a corrected amount, a different token) returns the original payment
281
- rather than an error, so the new terms are silently dropped.
817
+ Every write request carries a `request_id`. Pass a stable value and reuse the
818
+ exact same one on every retry (a timeout, a network blip, a restart) and you end
819
+ up with one payment, not two — and you get the same answer each time. One contract
820
+ covers `submitPayment` and `createQuoteRequest` alike.
821
+
822
+ - Using this SDK, `request_id` is **required** on both builders
823
+ (`preparePaymentRequest`, `prepareQuoteRequest`) and neither invents one — an id
824
+ the library chose is an id you cannot reuse on a retry. On the payment path it is
825
+ the only thing you manage: the SDK derives the on-chain deposit nonce
826
+ deterministically from it, so retries line up on their own.
827
+ - On the CLIs, `--request-id` is optional; when omitted an id is generated and
828
+ printed, so you can re-run with `--request-id <id>` to re-attempt the same
829
+ payment or recover the same batch. A bare re-run, with no `--request-id`, is a
830
+ **separate** payment or batch.
831
+ - Re-sending an id returns what it created the first time — the original payment,
832
+ or the original quote-request batch — and creates nothing new.
833
+ - A genuinely new request needs a **new** `request_id`. Reusing one with different
834
+ economics (a corrected amount, a different token) is rejected with `409`
835
+ `IDEMPOTENCY_TERMS_MISMATCH` — one id identifies one payment, or one batch.
836
+ - Deadlines and the signature are exempt, so a re-submission of the original
837
+ bytes is fine even though its deadlines have expired by then. This is what makes
838
+ a lost `createQuoteRequest` response recoverable: re-send it and you get your
839
+ `quote_request_id` back.
840
+ - `idempotent_replay` on the response tells you whether you were handed something
841
+ that already existed.
842
+ - `pending` is not success. It means accepted and still settling: re-submit the
843
+ same `request_id` to collect the result. A `failed` payment is terminal and
844
+ keeps its id, so recovering from one needs a new id.
282
845
 
283
846
  See [docs/idempotency.md](docs/idempotency.md) for the full model: the two
284
- guarantees and the per-chain differences (EVM/Tron vs Solana).
847
+ guarantees and which endpoints each covers, when the `409` applies, how to read
848
+ the outcome, and the per-chain differences (EVM/Tron vs Solana).
285
849
 
286
850
  ## What is in the package
287
851
 
288
852
  - `dist/` — compiled JavaScript and TypeScript declarations, ready to use
289
- - `dist/cli/` — the four command-line tools
853
+ - `dist/cli/` — the five command-line tools
290
854
  - `docs/` — reference documentation, including [idempotency and retries](docs/idempotency.md)
291
855
  - `LICENSE`, `THIRD-PARTY-NOTICES.txt` — license terms and third-party attribution