frontpage.
newsnewestaskshowjobs

Made with ♥ by @iamnishanth

Open Source @Github

Show HN: New Ensō – first public beta

https://untested.sonnet.io/notes/new-enso-first-public-beta/
231•rpastuszak•18h ago•81 comments

Show HN: A continuation of IRS Direct File that can be self-hosted

https://github.com/openfiletax/openfile
22•elijahwright_•7h ago•0 comments

Show HN: TokenDagger – A tokenizer faster than OpenAI's Tiktoken

https://github.com/M4THYOU/TokenDagger
254•matthewolfe•17h ago•70 comments

Show HN: Open-Source International Space Station Tracker ESP32/Arduino for $20

https://github.com/GuitarML/SpaceStationTracker
64•keyth72•3d ago•17 comments

Show HN: Local LLM Notepad – run a GPT-style model from a USB stick

https://github.com/runzhouye/Local_LLM_Notepad
13•davidye324•5h ago•2 comments

Show HN: Crush Check – AI relationship text analyzer

https://crushcheck.app
3•omarfarooq360•3h ago•0 comments

Show HN: Praxos – Context Management for AI Agents

9•mogusian•5h ago•3 comments

Show HN: Private real-time dictation app for Mac

https://github.com/aviaryan/Very-Fast-Dictation
12•aviaryan•14h ago•4 comments

Show HN: Octelium – FOSS Alternative to Teleport, Cloudflare, Tailscale, Ngrok

https://github.com/octelium/octelium
335•geoctl•1d ago•145 comments

Show HN: Timezone converter that tells you if your meeting time sucks

https://timezig.com
2•skrid•7h ago•0 comments

Show HN: I built a daily sunlight tracker

https://www.lumehealth.io/products
40•vickipow•4d ago•31 comments

Show HN: C.O.R.E – Opensource, user owned, shareable memory for Claude, Cursor

https://github.com/RedPlanetHQ/core
5•Manik_agg•8h ago•0 comments

Show HN: I'm an airline pilot – I built interactive graphs/globes of my flights

https://jameshard.ing/pilot
1509•jamesharding•3d ago•196 comments

Show HN: Audiopipe – Pipeline for audio diarization, denoising and transcription

https://github.com/nullwiz/audiopipe
4•nullwiz•10h ago•0 comments

Show HN: Attach Gateway – one-command OIDC/DID auth for local LLMs

https://github.com/attach-dev/attach-gateway
4•hammadtariq•11h ago•0 comments

Show HN: We're two coffee nerds who built an AI app to track beans and recipes

https://beanbook.app
51•rokeyzhang•10h ago•40 comments

Show HN: Summle – A little maths Game

https://summle.net
25•kirchhoff•4d ago•9 comments

Show HN: ArcFont – Font Embedding Model

https://github.com/JErnestoMtz/ArcFont
2•jernestomg•12h ago•0 comments

Show HN: Sharpe Ratio Calculation Tool

https://www.fundratios.com/
22•navquant•1d ago•16 comments

Show HN: QuizKnit, an open source quiz creator

https://quizknit.com
3•jibolash•12h ago•0 comments

Show HN: AGL a toy language that compiles to Go

https://github.com/alaingilbert/agl
83•alain_gilbert•5d ago•12 comments

Show HN: Vet – A tool for safely running remote shell scripts

https://getvet.sh
80•a10r•2d ago•37 comments

Show HN: MailMap – Turn emails into interactive stories on Google Maps

https://mailmap.site
3•Premananda•14h ago•0 comments

Show HN: Sink – Sync any directory with any device on your local network

https://github.com/sirbread/sink
144•sirbread•3d ago•89 comments

Show HN: Semantic-dictionary – A Python dictionary with semantic lookup

https://github.com/eu90h/semantic-dictionary
4•eu90h•15h ago•0 comments

Show HN: A tool to benchmark LLM APIs (OpenAI, Claude, local/self-hosted)

https://llmapitest.com/
52•mrqjr•1d ago•9 comments

Show HN: DotnetEbpf - Write Linux eBPF kernel applications in C#

https://github.com/KallDrexx/DotnetEbpf
4•KallDrexx•16h ago•0 comments

Show HN: Ravana, Multi AI Assistant Browser Desktop App

https://github.com/yogirk/ravana
3•ashtavakra•16h ago•1 comments

Show HN: A different kind of AI Video generation

44•fcpguru•5d ago•19 comments

Show HN: My Cross-Platform MySQL Parser

https://abbychau.github.io/mysql-parser/
4•abbychau•17h ago•0 comments
Open in hackernews

Show HN: Praxos – Context Management for AI Agents

9•mogusian•5h ago
Hey HN! We're Lucas and Soheil, the founders of Praxos (https://praxos.ai). Praxos is a context manager for AI Agents, providing everything you need to build stateful agents that don't break in production.

Praxos can parse any data source, from unstructured PDFs and API streams to conversational messages, to structured databases, and transform them into a single Knowledge Graph. Everything in this graph is semantically typed and its relationships are made explicit, turning data into a clean, queryable universe of understanding that AI can use without making mistakes.

Whether you need to query for the answer to a question or to extract data in a way that makes sense for the current use case, Praxos does it all, with no requerying needed. This enables AI apps to parse data end-to-end, and then act on it to deliver outputs across single-chain and multi-chain reasoning steps. Intermediate, final, and user-edited outputs can be added back to the knowledge graph, allowing Praxos to learn on the fly.

When we were building in insurance, we often ran into two major problems deploying AI:

First, LLMs would prove incapable of parsing documents such as property schedules and insurance policies. For reference, a property schedule may be a 50-page collection of Word, Excel, and PDF documents detailing construction, usage, and geographical information about a collection of physical properties. Recreating one object (a property) would mean combing through the files establish semantic, conceptual, spatial, and sometimes implicit linkages between the data.

The outcome: relationship information would be lost, left blank, or hallucinated.

Second, repeated calls to search, retrieve, and update information would sometimes lead to cascading errors. This became more frequent across complex tasks such as reading a document, fetching previous user information, performing a calculation, storing it, and then presenting it to the user.

We realized that for AI to deliver more useful and accurate responses that correctly use relationships in the document, these relationships need to be made explicit. Much of the contextual information is represented without the usage of words. In turn, this means that we cannot directly interact with them programmatically, and LLMs are forced to interpret them themselves, every single time. That’s when we started building Praxos.

We've set up a self-serve option with a free tier (up to a data cap) for hobbyists and early-adopters. For context (no pun intended), this should cover you for up to 200 document pages. You can register here: https://data.praxos.ai/sign-up.

Our first version is an SDK meant to cover you across all your data extraction, retrieval, and update needs.

Here's how it works:

Organizing information: Praxos sorts information into ontologies, which are structured schemas for storing data. These allow you to introduce predefined types, attributes, and relationships that guide how the knowledge graph is built and interpreted.

Processing input data: Praxos can handle any data source, ranging from PDFs to tabular data, JSONs, and dialog-like exchanges. Extraction is performed end-to-end. You don't need to OCR, chunk, or pre-process your inputs. Processing is as simple as passing in your file and selecting an ontology.

Retrieving information / memories: For each query, Praxos searches and retrieves related stored information by leveraging a combination of graph traversal techniques, vector similarity and key-value lookups. Search objects will return both the entities/their connections, as well as a sentence.

We’d love to hear what you think! Please feel free to dive in, and share any thoughts or suggestions with us over Discord (https://discord.gg/wVmrXD2dJA).

Your feedback will help shape where we take Praxos from here!

Comments

alakep•3h ago
Insurance is truly the place to find information challenges. It’s abstract but hypothetically structured well.

Seems like there are a lot of use cases for this.

soheils9•3h ago
Only hypothetically structured well. In truth, basically every carrier structures their data just a bit differently, and then on top of it, few insurance packages in commercial insurance policies from one carrier. and then you have to deal with appraisals, tech specs, and all the other external data that is basically randomly structured.
mogusian•3h ago
As well as pesky websites with either zero or crappy API support. You end up biting the bullet and entering the data by hand or taking a screenshot and hoping OCR does the job.