Ask a grounded question
Providers chosen server-side. Different-provider adversarial validation is always on.
Web findings are validated through the same pipeline as your sources.
Job
Pipeline running. Polling…
validated claim(s),
rejected.
· tokens
in /
out
· cost ¢
Response
Show byte-locked grounded answer (audit form)
Citations
-
[]
hash:
Per-claim validation detail ()
✓
grounded by
· adversarial challenged by
source:
·
Failed
Saved sources
| Name | Type | Tags | Status | Size | Hash | Created | |
|---|---|---|---|---|---|---|---|
|
—
|
|||||||
| No sources yet. | |||||||
Requests log
| Submitted | Objective | Status | |
|---|---|---|---|
| open | |||
| No requests yet. | |||
API keys
Every key is bound to your organization
()
and can only be used to act on its behalf.
| Name | Organization | Prefix | Scopes | Last used | Created | |
|---|---|---|---|---|---|---|
| revoked | ||||||
| No API keys yet. | ||||||
–
of
page of
Organization administration
Manage your team for .
AI model selection is controlled by Anchor platform staff and is
not user-configurable.
Owners can promote/demote team members and disable user accounts.
Admins and members can submit jobs and use shared resources but
cannot modify the team roster.
| Display name | Role | Status | Joined | ||
|---|---|---|---|---|---|
| active disabled | |||||
| No users yet. | |||||
Platform administration
Site-wide controls. Set the default AI model for each pipeline
stage, then optionally override it for individual organizations.
Default Markup
Decimal multiplier on raw provider cost (e.g.
0.01 = 1% markup,
0.5 = 50%). Applied to
every org unless they have a per-org override.
Current:
(%)
These are the models every organization inherits unless you've
set a per-org override. Each row is one stage of the grounding
pipeline; role tells you whether
the stage runs on the primary or adversarial provider.
| Pipeline stage | Slot | Provider | Model | Source | |
|---|---|---|---|---|---|
| Loading… | |||||
/
plan
claims used / quota
/
jobs (30d)
users
Override the system default for any stage of this org's
pipeline. Changes apply to the next job they submit.
| Pipeline stage | Slot | Provider | Model | Source | |
|---|---|---|---|---|---|
org override
this org has a pinned model.
system default
inheriting the platform default.
fallback
neither configured — runner will use a hard-coded default.
| Organization | Plan | Markup | Users | Jobs (this month) | Overrides | Joined | |
|---|---|---|---|---|---|---|---|
|
archived
|
default | ||||||
| Loading… | |||||||
Every API key issued on the platform, with the organization
that owns it. Use search to find a key by name, prefix, or
owning organization.
| Name | Organization | Prefix | Scopes | Last used | Created | Status | |
|---|---|---|---|---|---|---|---|
| revoked active | |||||||
| No API keys match. | |||||||
–
of
page of
Public, open-data REST APIs the grounding pipeline can
dynamically consult during a job. The router LLM picks
from this catalog by reading each entry's
description, so write
descriptions that say when to use the source.
See docs/public-sources.md
for the full schema.
| Slug / name | Description | Tags | Auth | Rate rpm | Status | |
|---|---|---|---|---|---|---|
| No public sources match. Use + New source above, or run scripts/seed_public_sources.py to import the starter catalog. | ||||||
–
of
page of
Default markup:
(%)
Click an organization to see its cost broken down by provider and model.
| Organization | Markup | Total cost | Total + markup | Net |
|---|---|---|---|---|
|
archived
|
||||
| No organizations. | ||||
| Loading… | ||||
| Totals | ||||
→
jobs ·
markup
Loading breakdown…
| Provider | Model | Stages | Input tokens | Output tokens | Raw cost | + Markup | Net |
|---|---|---|---|---|---|---|---|
| No billable stages in this window. | |||||||
| Totals | |||||||
Every API request the platform has served. Each row identifies
the originating organization and the API key (or user session)
that issued it. Filter by organization and/or API key, then
page through the matching results.
Pick an organization first.
| When | Organization | API key | Method | Endpoint | Status | Time | Request id |
|---|---|---|---|---|---|---|---|
|
|
|||||||
| No requests match the current filters. | |||||||
Showing
–
of
0 results
Page
of
Users with the platform-admin flag can configure AI models for
every organization on the platform. The bootstrap admin is set
via the ANCHOR_PLATFORM_ADMIN_EMAILS
environment variable.
| Display name | Org | Org role | ||
|---|---|---|---|---|
| (you) | ||||
| No platform admins. | ||||