Reference
API reference
Every client-facing endpoint of the IA4AI engine, with the role it needs, its parameters and body fields, generated from the engine itself.
#Basics
Base URL: https://ai.odux.io/app. Every path below is relative to it. (https://ai.odux.io on its own is the
website: a JSON call there gets a web page back.)
Authentication: send your workspace key in the x-api-key header, or as
Authorization: Bearer YOUR_IA_KEY. Everything you can reach is limited to your own workspace.
Roles: each endpoint names the role it needs. Your workspace key and your team's signed-in sessions have both
viewer (read) and investigator (write usage and settings). Turning content mode on or off, erasing captured
content and granting support access need your signed-in workspace admin; the workspace key can read content mode's
status and export what was captured. The extension key has only browser.
Approval code: three actions (export captured content, replace the key, change provider keys) also need your
workspace's approval code in the x-ia-approval header, once you have set one.
Without it they answer 403 with x-ia-approval: required; after five wrong tries, 429 for 15 minutes.
Errors: a JSON body with a detail that says what went wrong. 401 means no valid key, 403 a key without the
role, 404 nothing of yours by that id, 409 a conflict (such as a verification over budget), 422 a body that
does not validate, and 429 too many requests (retry shortly).
Dates: periods are UTC days. days counts back from now; start and end (YYYY-MM-DD, both included) pick a
range instead.
Parameters and body fields on this page are read from the engine's own schema each time these docs are built, so they match what the server accepts.
#Your workspace
Who a key belongs to, the state of each connection, and the workspace's own settings.
#Who am I
GET/whoami
The workspace and roles of the key or session you send, and what it can see.
Needs: a signed-in workspace admin.
curl -s "https://ai.odux.io/app/whoami" \
-H "x-api-key: YOUR_IA_KEY"#What is my AI doing
GET/workspace/summary
Usage and spend, the agents and models seen, which connections are live, the problems found (each with its fix) and the next step. The MCP tool my_ai_overview returns the same.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/summary" \
-H "x-api-key: YOUR_IA_KEY"#Connection status
GET/workspace/status
For each way of connecting, when IA4AI last received something from it. The Connect page's Check my connection button reads this.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/status" \
-H "x-api-key: YOUR_IA_KEY"#Access
GET/workspace/access
Whether the workspace has a key yet, and how it is reached.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/access" \
-H "x-api-key: YOUR_IA_KEY"#Issue the workspace key
POST/workspace/key
Issues the key your agents and tools use. It is shown once; IA4AI keeps only its hash. A workspace that already has a key keeps it unless you send replace: true, and then the old key stops working at once. Replacing needs the approval code.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
replace |
boolean | false |
curl -s -X POST "https://ai.odux.io/app/workspace/key" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Set the approval code
POST/workspace/approval-code
Sets the workspace's approval code, or changes it (current is the code in use). At least 8 characters. Only a fingerprint is stored: it cannot be read back.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
code (required) |
string | |
current |
string | `` |
curl -s -X POST "https://ai.odux.io/app/workspace/approval-code" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Contact address
PUT/workspace/contact-email
Set the address IA4AI writes to about this workspace.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
email (required) |
string |
curl -s -X PUT "https://ai.odux.io/app/workspace/contact-email" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Spend threshold
PUT/workspace/threshold
The 30-day spend the plan is measured against. Send null to go back to the default.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
threshold_usd |
number, optional |
curl -s -X PUT "https://ai.odux.io/app/workspace/threshold" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Privacy choice and seat prices
PUT/workspace/alternative-settings
The inputs to the own-hardware plan: privacy (no_preference, prefer_local or must_stay_local) and seat_prices, what you pay a month per seat of each chat app. A field left out keeps its value.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
privacy |
string, optional | |
seat_prices |
object, optional |
curl -s -X PUT "https://ai.odux.io/app/workspace/alternative-settings" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Provider keys and subscriptions
The provider keys the gateway calls with, and the read-only vendor admin connections.
#List provider keys
GET/workspace/providers
The stored provider entries and the models each may be used for. Keys are never returned.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/providers" \
-H "x-api-key: YOUR_IA_KEY"#Save provider keys
PUT/workspace/providers
Replace the workspace's provider entries. Each is a provider, the models it may be used for (empty means every model that provider serves; a prefix such as deepseek- allows a family), and the key. Entries are tried in order. Needs the approval code.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
providers (required) |
list of ProviderInput |
curl -s -X PUT "https://ai.odux.io/app/workspace/providers" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Test a provider key
POST/workspace/providers/test
Checks a key against the provider's /models endpoint. No tokens are spent.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
provider (required) |
string | |
key |
string | `` |
curl -s -X POST "https://ai.odux.io/app/workspace/providers/test" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#List vendor connections
GET/workspace/vendors
The connected vendor admin accounts (Claude Enterprise, Microsoft 365 Copilot) and their last sync.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/vendors" \
-H "x-api-key: YOUR_IA_KEY"#Connect a vendor
PUT/workspace/vendors/{vendor}
Tests the vendor's admin credentials, stores them encrypted, and pulls the last 30 days in the background. Read-only: IA4AI can read usage reports, nothing else.
Needs: investigator.
| Parameter | In | Type | Default |
|---|---|---|---|
vendor (required) |
path | string |
| Body field | Type | Default |
|---|---|---|
credentials (required) |
object |
curl -s -X PUT "https://ai.odux.io/app/workspace/vendors/<vendor>" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Sync a vendor now
POST/workspace/vendors/{vendor}/sync
Pull the vendor's latest usage now.
Needs: investigator.
| Parameter | In | Type | Default |
|---|---|---|---|
vendor (required) |
path | string |
curl -s -X POST "https://ai.odux.io/app/workspace/vendors/<vendor>/sync" \
-H "x-api-key: YOUR_IA_KEY"#Disconnect a vendor
DELETE/workspace/vendors/{vendor}
Forgets the credentials. Usage already pulled stays until the workspace's data is deleted.
Needs: investigator.
| Parameter | In | Type | Default |
|---|---|---|---|
vendor (required) |
path | string |
curl -s -X DELETE "https://ai.odux.io/app/workspace/vendors/<vendor>" \
-H "x-api-key: YOUR_IA_KEY"#Upload a provider bill
POST/workspace/provider-bill
A provider's usage export (CSV), so IA4AI's totals can be checked against the bill day by day.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
provider (required) |
string | |
csv (required) |
string |
curl -s -X POST "https://ai.odux.io/app/workspace/provider-bill" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Stored bills
GET/workspace/provider-bill
Per provider: the days covered and the amount.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/provider-bill" \
-H "x-api-key: YOUR_IA_KEY"#Reports
Every report reads call metadata only, never prompts. Most take days (the last N days) or a start and end date (UTC days, both included).
#Spend
GET/report/spend
Total spend and tokens, by model, by task type and by day, and the calls whose model has no known price yet.
Needs: viewer.
curl -s "https://ai.odux.io/app/report/spend" \
-H "x-api-key: YOUR_IA_KEY"#Spend per day
GET/report/daily
Cost and calls for every UTC day of the period (at most 92 days).
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 92 |
curl -s "https://ai.odux.io/app/report/daily" \
-H "x-api-key: YOUR_IA_KEY"#Agents
GET/report/agents
Each agent: model calls, cost, models, tools used, tool failures, and when it was last seen.
Needs: viewer.
curl -s "https://ai.odux.io/app/report/agents" \
-H "x-api-key: YOUR_IA_KEY"#What each agent does
GET/report/agent-work
Per agent and kind of job: the cost split (uncached input, cached input, output), context size and its growth, cache share, cost per task run, peak and off-peak spend, tools, and advice marked measured or estimated.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/agent-work" \
-H "x-api-key: YOUR_IA_KEY"#Context events
GET/report/context-events
Where each agent's uncached input comes from: new content, task starts, context resets and cache misses, with what each cost.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/context-events" \
-H "x-api-key: YOUR_IA_KEY"#Runs
GET/report/runs
Each agent's runs as numbered steps (time, tools, tokens, cost, cache, flags), with the candidate findings. Pass run_id to expand one run; summary=1 for the short form.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional | |
run_id |
query | string, optional | |
summary |
query | boolean | false |
curl -s "https://ai.odux.io/app/report/runs" \
-H "x-api-key: YOUR_IA_KEY"#Jobs worth a look
GET/report/job-review
Short summaries of the few jobs where something could change; every other job in one line. No saving is stated.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/job-review" \
-H "x-api-key: YOUR_IA_KEY"#One job
GET/report/job/{run_id}
One job written up: what was asked, how long it took, what it cost, how it ended, and every step grouped into pieces of work.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
run_id (required) |
path | string | |
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/job/<run_id>" \
-H "x-api-key: YOUR_IA_KEY"#Kinds of job
GET/report/jobs
Kinds of job ranked by cost, and tasks that were run again.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
curl -s "https://ai.odux.io/app/report/jobs" \
-H "x-api-key: YOUR_IA_KEY"#Cheaper-model tests
GET/report/model-tests
For each pair of current model and cheaper candidate: how many live calls the side-by-side test sampled, how many matched, and whether a key's model scope blocks the test.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
check |
query | string, optional |
curl -s "https://ai.odux.io/app/report/model-tests" \
-H "x-api-key: YOUR_IA_KEY"#Should you buy a local box
GET/report/local-box
Which calls a 128 GB local AI box could run, how long each would take, and the box's price in months of what those calls cost now. Never a saving until a replay passes.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/local-box" \
-H "x-api-key: YOUR_IA_KEY"#Own hardware
GET/report/alternative-plan
Your bill (API calls and chat seats) against open models on your own hardware in three tiers: the work each tier could take, the hardware that volume needs, and each mix's monthly cost rented or bought. privacy previews another choice without saving it. Never a saving until a replay passes.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional | |
privacy |
query | string, optional |
curl -s "https://ai.odux.io/app/report/alternative-plan" \
-H "x-api-key: YOUR_IA_KEY"#The plan
GET/report/plan
30-day spend against the threshold, and what each way of spending less covers.
Needs: viewer.
curl -s "https://ai.odux.io/app/report/plan" \
-H "x-api-key: YOUR_IA_KEY"#Recommendations
GET/report/recommendations
Cheaper models to test, and the swaps a verification run proved on your own traffic (evidence.verified_swaps).
Needs: viewer.
curl -s "https://ai.odux.io/app/report/recommendations" \
-H "x-api-key: YOUR_IA_KEY"#Bill check
GET/report/bill-check
IA4AI's totals against the uploaded bills, day by day, with a verdict.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/bill-check" \
-H "x-api-key: YOUR_IA_KEY"#Claude Code
GET/report/claude-code
What Claude Code did, from its own telemetry: sessions, prompts, requests, tools run and failed, errors.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/claude-code" \
-H "x-api-key: YOUR_IA_KEY"#Chat apps
GET/report/chat
Who uses which AI chat app: exact from vendor admin data, observed from the browser extension.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
curl -s "https://ai.odux.io/app/report/chat" \
-H "x-api-key: YOUR_IA_KEY"#Browser usage
GET/report/browser
From the extension: per tool, per person, people on two or more tools, and light users.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
curl -s "https://ai.odux.io/app/report/browser" \
-H "x-api-key: YOUR_IA_KEY"#Policy guard
GET/report/policy-guard
Counts of sensitive material (secrets, email, phone, card, confidential) per detector, agent and provider. Counts only.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
days |
query | integer | 30 |
start |
query | string, optional | |
end |
query | string, optional |
curl -s "https://ai.odux.io/app/report/policy-guard" \
-H "x-api-key: YOUR_IA_KEY"#Investigations
An investigation covers one period and ends in findings, conclusions and one next step.
#Open an investigation
POST/workspace/client-cases
On the paid plan, opens and runs an investigation of your own workspace for the last days days, or for a start to end range.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
days |
integer | 30 |
start |
string, optional | |
end |
string, optional | |
note |
string, optional |
curl -s -X POST "https://ai.odux.io/app/workspace/client-cases" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#List investigations
GET/workspace/client-cases
Every investigation of this workspace: period, state and finding count.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/client-cases" \
-H "x-api-key: YOUR_IA_KEY"#Get an investigation
GET/workspace/client-cases/{case_id}
One investigation in full: its findings, notes and reactions.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/workspace/client-cases/<case_id>" \
-H "x-api-key: YOUR_IA_KEY"#Conclusions
GET/workspace/client-cases/{case_id}/conclusions
The answer in brief: the period, the totals, up to five conclusions, the next step and the chart's data. Every number in it is from the findings.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/workspace/client-cases/<case_id>/conclusions" \
-H "x-api-key: YOUR_IA_KEY"#React to a finding
POST/workspace/client-cases/{case_id}/react
Accept a finding, reject it, or ask a question about it (kind: accept, reject or question). A reaction never starts a new run by itself; an analyst picks it up.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
| Body field | Type | Default |
|---|---|---|
finding_index (required) |
integer | |
kind (required) |
string | |
note |
string, optional |
curl -s -X POST "https://ai.odux.io/app/workspace/client-cases/<case_id>/react" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Approve or decline a verification
POST/workspace/client-cases/{case_id}/verification
Approve (it runs on your own provider keys, within your daily budget) or decline the verification an analyst proposed.
Needs: investigator.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
| Body field | Type | Default |
|---|---|---|
approve (required) |
boolean |
curl -s -X POST "https://ai.odux.io/app/workspace/client-cases/<case_id>/verification" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Outcome
GET/workspace/client-cases/{case_id}/outcome
After a verified swap: whether you switched, whether the saving happened, and whether quality held.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/workspace/client-cases/<case_id>/outcome" \
-H "x-api-key: YOUR_IA_KEY"#Runs, cases and traces
A case is one analysed agent run. These are what ingest/trace produces.
#List cases
GET/cases
Every case this key may see: case id, job, finding and verdict counts, cost.
Needs: viewer.
curl -s "https://ai.odux.io/app/cases" \
-H "x-api-key: YOUR_IA_KEY"#Get a case
GET/case/{case_id}
One case file: findings, verdicts, cost and root cause.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/case/<case_id>" \
-H "x-api-key: YOUR_IA_KEY"#A case's trace
GET/case/{case_id}/trace
The trace a case was built from.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/case/<case_id>/trace" \
-H "x-api-key: YOUR_IA_KEY"#Cheaper models for a case
GET/case/{case_id}/cost-cuts
Cheaper models that meet the job's needs, as candidates to test. A saving is stated only for a swap a verification run proved.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
curl -s "https://ai.odux.io/app/case/<case_id>/cost-cuts" \
-H "x-api-key: YOUR_IA_KEY"#Get a trace
GET/trace/{trace_id}
One trace. When content was captured it shows that it was (roles, counts, hashes), never the prompt.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
trace_id (required) |
path | string |
curl -s "https://ai.odux.io/app/trace/<trace_id>" \
-H "x-api-key: YOUR_IA_KEY"#Compare two runs
GET/diff/{case_a}/{case_b}
Which decisions, tools, models and cost changed between two cases.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_a (required) |
path | string | |
case_b (required) |
path | string |
curl -s "https://ai.odux.io/app/diff/<case_a>/<case_b>" \
-H "x-api-key: YOUR_IA_KEY"#Sending usage
Write routes: they need the investigator role, which your workspace key has. The OTLP routes take Protobuf or JSON, picked from the Content-Type.
#Send a run
POST/ingest/trace
One agent run: a trace (its spans), and optionally a job. With a job a case is built; without one the trace is stored for later.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
job |
Job, optional | |
trace (required) |
object or list of any |
curl -s -X POST "https://ai.odux.io/app/ingest/trace" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#OTLP traces
POST/v1/traces
OpenTelemetry traces. /ingest/otlp is the same route.
Needs: investigator.
curl -s -X POST "https://ai.odux.io/app/v1/traces" \
-H "x-api-key: YOUR_IA_KEY"#OTLP metrics
POST/v1/metrics
OpenTelemetry metrics. /ingest/metrics is the same route.
Needs: investigator.
curl -s -X POST "https://ai.odux.io/app/v1/metrics" \
-H "x-api-key: YOUR_IA_KEY"#OTLP logs
POST/v1/logs
OpenTelemetry logs. /ingest/logs is the same route.
Needs: investigator.
curl -s -X POST "https://ai.odux.io/app/v1/logs" \
-H "x-api-key: YOUR_IA_KEY"#Golden sets and verification
Only for a workspace with captured content. See Capture, golden sets and verification.
#Build a golden set
POST/golden-sets
Stores a set of your captured inputs with the answers the current model gave. Building again from unchanged traffic returns the existing set.
Needs: investigator.
| Body field | Type | Default |
|---|---|---|
tenant_id |
string, optional | |
job_id |
string, optional | |
max_cases |
integer | 100 |
description |
string, optional |
curl -s -X POST "https://ai.odux.io/app/golden-sets" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#List golden sets
GET/golden-sets
Metadata only: version, counts, sources, dates. Never content.
Needs: viewer.
curl -s "https://ai.odux.io/app/golden-sets" \
-H "x-api-key: YOUR_IA_KEY"#Get a golden set
GET/golden-sets/{set_id}
The set's cases. Every read is audited.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
set_id (required) |
path | string |
curl -s "https://ai.odux.io/app/golden-sets/<set_id>" \
-H "x-api-key: YOUR_IA_KEY"#Verify a case
POST/case/{case_id}/verify
Replays a golden set on the current model and each cheaper candidate, in the background, with your own provider keys. Refused with 409 when the worst-case cost exceeds what is left of today's budget.
Needs: investigator.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id (required) |
path | string |
| Body field | Type | Default |
|---|---|---|
set_id (required) |
string | |
judge_model |
string, optional | |
seeds |
integer | 3 |
curl -s -X POST "https://ai.odux.io/app/case/<case_id>/verify" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#List verification runs
GET/verification-runs
Your verification runs; filter with case_id.
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
case_id |
query | string, optional |
curl -s "https://ai.odux.io/app/verification-runs" \
-H "x-api-key: YOUR_IA_KEY"#Get a verification run
GET/verification-runs/{run_id}
A run's status and, once complete, its report: swaps (verified) and unproven (ties).
Needs: viewer.
| Parameter | In | Type | Default |
|---|---|---|---|
run_id (required) |
path | string |
curl -s "https://ai.odux.io/app/verification-runs/<run_id>" \
-H "x-api-key: YOUR_IA_KEY"#Content mode
Off unless your workspace's own admin turns it on. Changing anything here needs that admin, signed in; the status and the export can also be read with the workspace key.
#Content mode status
GET/workspace/content-consent
Whether content mode is on, what it covers and when it ends.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/content-consent" \
-H "x-api-key: YOUR_IA_KEY"#Turn content mode on
POST/workspace/content-consent
For 1 to 90 days, for the kinds you choose (tool_arguments, reasoning, prompts_answers), optionally for some agents only, with a purpose and the accepted terms. It closes by itself.
Needs: a signed-in workspace admin.
| Body field | Type | Default |
|---|---|---|
days (required) |
integer | |
kinds (required) |
list of string | |
scope_agents |
list of string | |
purpose |
string | `` |
terms_accepted |
boolean | false |
analysis_model |
string | none |
curl -s -X POST "https://ai.odux.io/app/workspace/content-consent" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Turn content mode off
POST/workspace/content-consent/revoke
Stops capture at once and records it.
Needs: a signed-in workspace admin.
curl -s -X POST "https://ai.odux.io/app/workspace/content-consent/revoke" \
-H "x-api-key: YOUR_IA_KEY"#Export captured content
GET/workspace/content-export
Everything captured for this workspace, decrypted, as JSON. Needs the approval code. Every read is recorded.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/content-export" \
-H "x-api-key: YOUR_IA_KEY"#Erase captured content
POST/workspace/content-erase
Removes every copy. Confirm by sending the workspace id; send stop: true to turn content mode off first.
Needs: a signed-in workspace admin.
| Body field | Type | Default |
|---|---|---|
confirm (required) |
string | |
stop |
boolean | false |
curl -s -X POST "https://ai.odux.io/app/workspace/content-erase" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Support access
GET/workspace/support-access
Whether IA4AI support may read captured content, until when, and every read made under it.
Needs: viewer.
curl -s "https://ai.odux.io/app/workspace/support-access" \
-H "x-api-key: YOUR_IA_KEY"#Grant support access
POST/workspace/support-access
Lets IA4AI support read captured content for 24 hours to 7 days, to help with a problem. Audited.
Needs: a signed-in workspace admin.
| Body field | Type | Default |
|---|---|---|
hours (required) |
integer |
curl -s -X POST "https://ai.odux.io/app/workspace/support-access" \
-H "x-api-key: YOUR_IA_KEY" -H 'content-type: application/json' -d '{…}'#Revoke support access
POST/workspace/support-access/revoke
Ends it at once. Audited.
Needs: a signed-in workspace admin.
curl -s -X POST "https://ai.odux.io/app/workspace/support-access/revoke" \
-H "x-api-key: YOUR_IA_KEY"#Browser extension
The only two routes the extension key (iax_…) can call.
#Create the extension key
POST/workspace/extension-key
Shown once. A new one disconnects every browser on the old one.
Needs: investigator.
curl -s -X POST "https://ai.odux.io/app/workspace/extension-key" \
-H "x-api-key: YOUR_IA_KEY"#Extension settings
GET/browser/config
The site rules and settings the extension fetches every hour.
Needs: browser (the extension key).
curl -s "https://ai.odux.io/app/browser/config" \
-H "x-api-key: YOUR_IA_KEY"#Send usage events
POST/browser/events
A batch of usage events. Fields outside the declared ones are dropped.
Needs: browser (the extension key).
curl -s -X POST "https://ai.odux.io/app/browser/events" \
-H "x-api-key: YOUR_IA_KEY"#Records and models
#Verify the evidence log
GET/audit/evidence-log/verify
Re-checks the hash chain and signatures and says whether any record was changed since it was written.
Needs: viewer.
curl -s "https://ai.odux.io/app/audit/evidence-log/verify" \
-H "x-api-key: YOUR_IA_KEY"#Model catalog
GET/catalog
Every model with its prices, capabilities and lifecycle, and how it did in your own verification runs.
Needs: viewer.
curl -s "https://ai.odux.io/app/catalog" \
-H "x-api-key: YOUR_IA_KEY"#Health
GET/health
200 when the engine is up. No key needed.
Needs: none.
curl -s "https://ai.odux.io/app/health"
