Release DevFeed 0.1.1 · abhi1693/devfeed.tech

#250 - Track source onboarding in a live pipeline with profile, assessment, branding, review, and ingestion stages, plus targeted retries and readiness checks before activation.
#251 - Manage article readiness through a live lifecycle board, reuse successful analysis steps after later failures, and publish only after enrichment, presentation, and image processing are ready.
#241 - Choose partner import intervals supported by the account tier, from daily imports to hourly synchronization.
#240 - Import larger partner catalogs with configurable page limits beyond the previous fixed caps.
#258, #260, #261, #262 - Let automated clients discover reader API documentation, the configured sign-in provider, OAuth resource metadata, and MCP server capabilities through public discovery URLs.
#235 - Audit failing publisher feeds for verified replacement candidates without changing source approval or ingestion state.
#277 - End the provider SSO session when signing out and honor provider logout across the reader, admin, partner portal, Chrome, and Edge.

#230, #236, #246 - Reduce database work for processing history, directory counts, topic retention, and inference reports while preserving exact totals.
#231 - Limit workload aggregation to active jobs and completed jobs within the selected reporting window.
#234, #239 - Reduce repeated visibility and filtering work in source filters and broad personalized recommendation queries.
#233 - Reuse source-tag identity lookups until the tag catalog changes.
#227 - Release completed analysis catalog indexes instead of retaining full generations in worker memory.

#229, #245 - Keep source, topic, and tag pages recoverable during temporary API failures, with navigation and retry available on web, Chrome, and Edge.
#278 - Explain failed or cancelled sign-in attempts and preserve the intended destination when retrying, including browser extension sign-in.
#249 - Attribute articles to their original publisher, distinguish original headlines from DevFeed previews, and publish article tags as search keywords.
#226, #247 - Show actionable partner import and assessment errors, and display rejected, paused, and withdrawn product statuses accurately.
#263, #276 - Keep Cloudflare's WebMCP bridge and authenticated reader navigation on configured LAN origins compatible with the script security policy.
#257 - Serve the public homepage as Markdown when requested through content negotiation or /index.md.
#256 - Publish crawler content-use preferences that permit search while declining AI training and AI input.
#248 - Use consistent DevFeed favicon and touch icons in the partner portal.

Quiesce source dispatch and drain older coupled source workers before upgrading the database through migration 0033. Start the updated APIs and source/image workers before resuming dispatch; older workers cannot consume the split source operations introduced in 0032.
Schedule migrations 0027–0031 with allowance for index creation and the inference-summary backfill, which blocks ledger writes while completing.
Register each OIDC client's back-channel logout URL and exact post-logout redirect URI, including configured reader aliases, as documented in browser OIDC logout. Existing reader, admin, and partner browser sessions must sign in again.
Deploy the updated APIs and reader before distributing browser extensions 0.1.22, which use coordinated provider sign-out and login recovery.
For MCP server-card discovery, deploy the updated MCP service first and configure the same DEVFEED_MCP_PUBLIC_URL on MCP and the reader. When the reader needs a private transport address, set its optional server-only DEVFEED_MCP_INTERNAL_URL.

Navigation Menu