HTTP API

Run code over HTTP

One submission in, one result out, in the standard submission format: the same fields, status ids and language ids code-execution clients already use. Authenticate with your API key as Authorization: Bearer … or X-Auth-Token: ….

Run and wait

shell

curl -s "https://boltzlabs.cloud/api/execute?wait=true" \
  -H "X-Auth-Token: $BOLTZLABS_API_KEY" -H "Content-Type: application/json" \
  -d '{"source_code": "print(int(input()) * 2)", "language_id": 100,
       "stdin": "21", "expected_output": "42", "cpu_time_limit": 1}'

response.json

{
  "stdout": "42\n",
  "stderr": null,
  "compile_output": null,
  "message": null,
  "status": { "id": 3, "description": "Accepted" },
  "time": "0.012",
  "memory": 9216,
  "token": "5d1e3c2a-…"
}

fields=* returns every field; base64_encoded=true takes and returns source, input and output as base64.

Request

Field	Default	Meaning
source_code	required	The program.
language_id	required	See the table below.
stdin	—	The program's input.
expected_output	—	Compare stdout to it: Accepted or Wrong Answer.
cpu_time_limit	15	Seconds of CPU the program may use.
cpu_extra_time	1	How far past the CPU limit it may run before it is stopped.
wall_time_limit	30	Seconds of wall time; with only a CPU limit, about 2× it + 1.
memory_limit	262144	KB.
redirect_stderr_to_stdout	false	Merge stderr into stdout.
supersede_key	—	A newer run with the same key cancels this one.

Limits only ever lower the platform's own; above them is a 422. So are fields this platform does not offer — compiler options, command-line arguments, extra files, callbacks and network access — rather than being silently ignored.

Response

status	{ id, description } — see the table below
stdout, stderr	The program's output
compile_output	The compiler's messages (compiled languages)
message	Why a run stopped, e.g. Time limit exceeded
time, wall_time	Seconds, as strings: "0.012". Sandbox setup is not in them.
memory	Peak memory, KB
exit_code, exit_signal	How the program ended
token	To fetch this submission again

Statuses

1	In Queue
2	Processing
3	Accepted
4	Wrong Answer
5	Time Limit Exceeded
6	Compilation Error
7	Runtime Error (SIGSEGV)
8	Runtime Error (SIGXFSZ)
9	Runtime Error (SIGFPE)
10	Runtime Error (SIGABRT)
11	Runtime Error (NZEC)
12	Runtime Error (Other) — including memory and output limits
13	Internal Error — our failure, never billed

Languages

GET /api/languages is the live list. Ids, codes and versions are listed on Supported languages.

Queue now, fetch later

Without wait=true you get a token at once; fetch the result with it. Results are kept for an hour.

shell

POST   /api/execute                 → { "token": "…" }
GET    /api/execute/{token}         → the submission (status 1 and 2 while it waits and runs)
DELETE /api/execute/{token}         → removes it, and cancels it if unfinished

Batches

Up to 20 submissions in one request — a problem's test cases, say. They run in parallel.

shell

POST /api/execute/batch            {"submissions": [ {…}, {…} ]}  → [{"token": "…"}, …]
GET  /api/execute/batch?tokens=a,b → {"submissions": [ {…}, {…} ]}   (in the order asked)