Proposal: Akash Data Indexing, Analytics, and Agent Integration w/ Chakra

[0xnirmal]

Proposal: Akash Data Indexing, Analytics, and Agent Integration w/ Chakra

Executive Summary

We propose a comprehensive data solution for Akash Network that will enhance data accessibility, provide unique insights into GPU utilization, and enable AI agents to have access to Akash fundamentals.

This proposal consists of three main components:

1. Specialized chain indexing
2. A dedicated analytics dashboard
3. Agent Kit integration

Background on Chakra

Chakra's mission is to organize the world's structured data.

Along those lines, we've spent the last 2 months building a data warehouse on top of decentralized storage and compute primitives, using technology like Akave and Akash.

• You can largely think about this as a plug-and-play replacement for Snowflake or BigQuery, but running on decentralized primitives.
• We have an extremely generous free-plan and expect most consumers will fall under this category.

The article below summarizes our masterplan and along the lines of creating a rich, open data marketplace on top of the data warehouse, we'd like to prioritize Akash data as one of our first datasets.

Structurally, this provides Akash with the following advantages:

• A trusted third-party provider of all indexed data.
• A home for analysts, investors, and researchers to query and answer questions about the growing utilization of the platform. The list of desired metrics from SIG Analytics forum continues to grow and we expect the community can step in to address some of these needs, once there is a common, unified data store.
  • https://github.com/akash-network/community/blob/main/sig-analytics/prd.md
• Another maintainer for the stats repository. On this note, our team has already proposed several updates to this repository:
  • feat(analytics): GPU Usage console#580
  • feat(docs): Indexer reference / Migration Clean-Up console#561
  • The Chakra team is now running its own indexer for Akash data (thanks to the help from Max)

1. Specialized Chain Indexing with the Chakra Data Warehouse

Current Challenge

While existing blockchain indexing solutions like Numia provide basic chain data, they fail to capture Akash-specific metrics that are crucial for understanding network utilization, particularly GPU resources.

This creates a gap in transparent, reliable data access for the Akash ecosystem.

While Akash Stats is a fabulous resource, today, researchers have to run their own indexer to answer questions that do not conform to the standards of the existing endpoints on the service.

Proposed Solution

Implement a specialized indexing solution through the Chakra Data Warehouse that will:

• Index all chain data with a focus on Akash-specific metrics
• Track detailed GPU-specific statistics from the non-standard Cosmos indexers
  • https://github.com/akash-network/console/blob/main/apps/indexer/src/indexers/akashStatsIndexer.ts
• Monitor provider performance and availability

Key Benefits

• Establishes a trusted third-party source of truth for Akash network data
• Provides granular insights into Akash specific metrics like GPU utilization and availability
• Enables data-driven decision making for network participants
• Supports network growth through transparent resource tracking

• Tables in scope:

2. Dedicated Analytics Dashboard (chakra.dev/akash)

Overview

We will create a dedicated analytics dashboard at chakra.dev/akash that provides comprehensive insights into Akash network metrics, with a special focus on GPU utilization.

Features

• Real-time GPU utilization metrics
• Provider performance analytics
• Network health indicators
• Fully configurable long-term, we will support additional asks on visualization / metrics.

Open Source Contribution

All queries and visualization components will be contributed back to the Akash core repository, enabling:

• Community-driven improvements
• Integration with other Akash tools
• Standardization of data analysis methods

3. Agent Kit Integration

Vision

Integrate Akash network data with Agent Kit to enable AI-powered network interaction and data discovery.

Implementation

• Provide agents with direct access to indexed Akash data
• Enable natural language queries for network statistics (we are currently building our Agent Kit)
• Build a Twitter agent specifically to cite Akash data usage and metrics (along with other DePINs - we are working with the folks at Depin.Ninja via DePIN NinjAI

Timeline and Deliverables

Phase 1: Data Indexing (Months 1-2)

• Productionize specialized indexing infrastructure
• Implement Akash-specific metrics tracking
• Establish data validation and quality controls

Phase 2: Dashboard Development (Months 2-3)

• Publicize chakra.dev/akash analytics platform
• Develop custom visualization components
• Contribute queries to Akash core repository

Phase 3: Agent Integration (Months 3-4)

• Implement Agent Kit data access
• Develop specialized agents
• Create documentation and examples

Success Metrics

• Complete indexing of all Akash-specific chain data
• Dashboard uptime and reliability > 99.9%
• Query response time <5 seconds (for common, acceptable query sizes)
• Successful integration of agent use cases

Next Steps

1. Technical review and feedback from Akash team
2. Finalize specific metrics and KPIs (if any for chakra.dev/akash)
3. Establish development milestones
4. Begin Phase 1 implementation

Budget

Monthly Plan

The proposed budget for ongoing services is $1,500 per month (or equivalent in AKT at market rate), which covers:

• Infrastructure costs for specialized chain indexing
• Data warehouse maintenance and operations
• Dashboard hosting and updates
• Ongoing technical support and monitoring

Notable, we are not charging any margin here and both the monthly and annual rate reflect costs for running infrastructure and development work - nothing materially more. The cost structure is designed to be sustainable while providing significant value to the Akash ecosystem.

Annual Plan

For upfront annual commitment:

• $16,200 total (equivalent to $1,350 per month)
• 10% discount compared to monthly billing
• Can be paid in USD or AKT (at market rate at time of payment)

[rodri-r]

Hey this looks interesting and I think budget and time frame is fair.

However most of this data seems like it is already covered in console and Akash stats, there is a way to run an indexer as well https://github.com/akash-network/console/tree/main/apps/indexer

I do like your dashboard and how it goes into more specifications about models being used. I also like the agent integration part. Nice job!

[0xnirmal]

Hey @rodri-r - thanks so much for the feedback! Glad to hear you think the proposal / budget / timeframe are fair!

Some thoughts:
most of this data seems like it is already covered in console and Akash stats -> We contributed the changes shown in our dashboard to the stats indexer 😄 akash-network/console#580

We did this as POC to show that all queries we write will get contributed back to open source, and it looks like there is demand for these types of queries. Do let us know if you have any other metrics you'd like prioritized and we can definitely support. We intend to use this as a launching point but will get feedback from other stakeholders.

there is a way to run an indexer as well -> Yup totally agree - the current indexer works great! We want to make it easier for developers and analysts who don't run an indexer to still have access to the raw / processed data (i.e. Dune).

Hope this makes sense - do let me know if any add'l feedback.

[rodri-r]

Nice, I did not know you had contributed those changes to the indexer as POC.

All makes sense to me.