Free tools Windows power users keep installed
One-click scans. No signup required.
CanonKeeper is a project submission that describes an AI-assisted way to reconcile fictional-universe lore when sources disagree, timelines shift, or an artifact’s custody depends on several linked events. Its author, Sandeep Chakravartty, says it uses Sanity CMS to structure lore records and a query-and-agent workflow to trace claims to supporting, conflicting, or superseding sources. These are the project’s described capabilities, not independently verified results.
What CanonKeeper is—and is not
CanonKeeper is presented as a Sanity developer-challenge project, not as a documented commercial product or an independently reviewed service. The project post describes an engine for resolving fictional canon questions that ordinary keyword search can struggle with: for example, whether a timeline detail was retconned, which conflicting account takes precedence, or how an object passed between characters.
As an Amazon Associate I earn from qualifying purchases.
The source is a DEV Community submission by Sandeep Chakravartty marked “Posted on Sep 30”; the year is not established in the retrieved text. The post is primary evidence for how its author describes the project, but it does not independently establish the code’s performance or whether the demo and repository are still available.
Recommended Free Tools
How the lore records are organized
The submission says CanonKeeper stores structured lore documents in Sanity CMS. Instead of treating lore as undifferentiated prose, its model represents claims, sources, characters, events, and locations as linked records.
#1 Best Overall
Claims and their status
A claim is described with subject-predicate-object fields, references to sources, timeline markers, a confidence score, and a canon status. The statuses listed are CANON, RETCONNED, DISPUTED, and AMBIGUOUS. Claims can also link to other claims as contradictions, corroborations, or supersessions.
Sources and authority
Source records are said to include a source type, edition, publication date, authority level, and references to sources they supersede. The author gives this authority order: PRIMARY_CANON > OFFICIAL_SECONDARY > REFERENCE > COMMUNITY. In the post’s example, a revised ledger with higher authority supersedes an earlier secondary account. That ordering is the project’s stated rule, not a universal standard for fictional canon.
Rank #2
- An illustrated guide to Dungeons & Dragons’ beloved fifth edition told through interviews, artwork, and visual ephemera from the designers, storytellers, and artists who bring it to life.
How a canon question is resolved
The described agent queries the Sanity dataset using GROQ and MCP. It identifies relevant entities and timeframe constraints, follows references between claims and sources, and expands contradiction and supersession relationships. In practical terms, the workflow is intended to move beyond matching words: it follows the records that explain why one account supports, conflicts with, or replaces another.
For an artifact-possession question, for example, the project says it can use multi-hop evidence deduction across connected claims and events rather than relying on one isolated mention. The submission also describes timeline-retcon detection. It does not provide independent validation or measured accuracy for either capability.
Rank #3
What the evidence graph is meant to show
The author describes an interactive SVG graph with nodes for a question, claims, sources, and a verdict. Its listed edge types include retrieves, supported_by, contradicts, supersedes, and resolves_to. If implemented as described, this graph would make the route from a question to its supporting evidence visible instead of presenting only a bare answer. The available source does not establish how the graph behaves in a live deployment.
What the project description establishes
- CanonKeeper is described as a project submission for a Sanity developer challenge.
- Sanity CMS is identified as the structured lore store; GROQ and MCP are identified as query integrations.
- The stated resolution approach uses source authority alongside links for contradiction, corroboration, and supersession.
- The author claims support for timeline-retcon detection, multi-hop evidence deduction, and an interactive evidence graph; these capabilities have not been independently validated here.
Implementation details and availability
The post mentions Sanity MCP integration, includes a GROQ query example, and names a Sanity project ID of demo-universe with a production dataset. It also refers to a seeding script using @sanity/client. A video walkthrough and a GitHub repository named Sanity-MCP-Hackathon are linked from the post, but their current contents and availability are not established by the retrieved material.
Quick Recap
Rank #4
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →
