Grok Thyself (grokthyself.com) is a tool developed by Jan_Wilmake and launched on April 7, 2026. It synchronises a user's X (Twitter) posts, bookmarks, and likes into a private GitHub repository of the user's choice, creating a structured, AI-queryable personal knowledge base from social media activity.[1]
The tool was inspired by Andrej Karpathy's viral thread on LLM knowledge bases (April 5, 2026), in which Karpathy advocated for using LLMs to build and maintain personal knowledge bases. Jan had been building personal knowledge infrastructure since 2024, and Karpathy's post provided the catalyst to productise the X-data sync component.[2]
Jan's initial question to his audience: "who wants a continuous sync of all their tweets+bookmarks+likes into a private github repo to augment their knowledge base?" — received strong interest.[2]
Within 24 hours of launch, Grok Thyself achieved:[4]
| Metric | Count |
|---|---|
| Likes on launch tweet | 185 |
| Bookmarks | 290 |
| Sign-ups | 40 |
| Paid subscribers | 3 |
Jan described the launch as "a success" and immediately pivoted to the next product in the pipeline: Self-Driving_Repos.[4]
By April 13, 2026, sign-ups had grown to 48 — but only 3 users were fully set up and paying, indicating a significant activation gap. Jan was actively evaluating the onboarding flow to close this gap, and the community was leaning toward a paid-only model (no free tier).[5]
An unexpected development: someone in the web3 community created a memecoin for Grok Thyself through a decentralised crowdfunding mechanism, motivated by a desire to support solo developers. Jan had no involvement in or awareness of this until notified by a user.[1]
Several sign-ups did not complete the setup flow, prompting Jan to individually reach out to approximately 14 users asking if they needed help — a characteristic pattern of hands-on user engagement.[1]
Grok Thyself's initial offering includes backfill of up to 1,000 historic bookmarks and 1,000 historic posts. On April 9, 2026, a user (@wgw_eth) requested backfill for approximately 30,000 posts. Jan proposed a pay-per-post model at approximately $0.005+ per post, reflecting the underlying X API costs. An alternative path — users manually exporting their X data archive and transforming it to markdown — was offered as a cheaper but more labour-intensive option.[6]
Grok Thyself is the data ingestion layer for Jan's broader personal knowledge base architecture:
The x-archive repository used as a source for this wiki was itself generated by an early version of the same sync pipeline.[1]
Grok Thyself is the productised evolution of the earlier socialdatapod concept, which focused on syncing full X post history to a personal database as an MCP server. Where socialdatapod was developer-focused (a GitHub repo + MCP), Grok Thyself packages the same idea for a broader audience with a web-based setup flow, GitHub App integration, and paid subscriptions.
Jan explored whether to host the sync process on the user's machine or in the cloud. He concluded that cloud-based continuous sync (via private GitHub repos) is superior because users should not need to keep their computer running. This preference motivated the subsequent development of Self-Driving_Repos.[1]
On April 17, 2026, X announced a pricing update: "Owned Reads Now $0.001 + Other Changes Effective April 20, 2026." Jan spent the morning (07:02–08:16) reading the email announcement, the X Developer Community thread, and using ContextArea to draft a response. He also explored X API timeline documentation (Timelines overview, Get Timeline, Quickstart) and revisited the socialdatapod/grokthyself codebase on uithub.[7]
This pricing change directly affects GrokThyself's cost model, as the tool reads X timelines via the API. The GOALS.md document identified resolving the pricing impact as urgent before April 20 — requiring an audit of how GrokThyself uses "Owned Reads" endpoints, a cost impact calculation per user per month, and possible pricing or rate-limit adjustments.[5]
Jan also explored X API timeline endpoints (09:27–09:36), suggesting he was investigating alternative approaches to data retrieval that might mitigate the cost increase.[7]
As of April 20, 2026, the new X API pricing is live. The $0.001 per owned read is now in effect. Some developers in the community reported potential exposure of $8,000/day for high-volume use cases. Jan's response strategy — drafted using ContextArea on April 17 — has not yet been publicly communicated to GrokThyself users. As of April 25, whether the cost audit was completed remains unconfirmed, and the GOALS.md document continues to flag this as an immediate priority — requiring a before/after cost comparison per user per month and possible pricing or rate-limit adjustments.[5]
On April 21, 2026, Jan read tessl.io's "The Context Flywheel: Why the Best AI Coding Teams Will Win on Context" — a framework arguing that context accumulation compounds like a flywheel. GrokThyself is literally building a personal context flywheel from X data: every sync adds more signal, every AI query over the archive is richer than the last. The GOALS.md document identified "context flywheel" as a high-priority vocabulary hook for GrokThyself's marketing copy — reframing the product from "X data backup" to "your personal context flywheel."[5] [8]
On April 24, 2026, Jan spent 37 seconds reading a tweet by Celeste Amadon: "I would pay anything for a CRM that tracks cross platform (texts, emails, dms) everyone I know and when I last talked with them." GrokThyself's architecture — GitHub-stored, structured Markdown, agent-queryable — could power exactly this use case. This angle expands GrokThyself's addressable market beyond X data backup into the broader personal data sovereignty space, positioning it as "your personal relationship memory, stored privately in GitHub, queryable by AI."
As of April 2026, Jan was actively researching competitors in the X data backup space, including searching "backup x data" on Reddit and monitoring tools like Conare.ai and @dabit3's xAI/OpenClaw daily report pipeline. The key competitor insight: storage was a solved problem, but retrieval and intelligence on top of the data was not yet addressed by existing tools.[5]
On April 17, Jan spent ~6.5 minutes reading Nader Dabit's thread on using the xAI API + @openclaw to build curated X feeds and daily email reports — a pattern that overlaps with GrokThyself's value proposition but approaches it from the curation/intelligence angle rather than raw data backup.[7]
x-archive/raw/x-posts/2026-04-07.md ↑x-archive/raw/x-posts/2026-04-05.md ↑x-archive/raw/x-posts/2026-04-12.md ↑x-archive/raw/x-posts/2026-04-08.md ↑x-archive/entries/GOALS.md ↑x-archive/raw/x-posts/2026-04-09.md ↑x-archive/raw/activity-summary/2026-04-17.md ↑x-archive/raw/browser-history/2026-04-21.md ↑