Sanatana Sampattiसनातनसम्पत्तिः

The Taittirīya Saṃhitā, Brāhmaṇa and Āraṇyaka, laid out by the traditional pañcāśat. The Saṃhitā has its padapāṭha throughout, Bhaṭṭa Bhāskara's and Sāyaṇa's bhāṣyas beside the text, and the ācārya's recording. Svarānugamī means "follows the svara": as the chant plays, the text moves with it, phrase by phrase and word by word.

• All 2,198 pañcāśats of the Saṃhitā carry their padapāṭha. The Brāhmaṇa (1,752 pañcāśats) and the Āraṇyaka (577) are there with their korvai.
• Bhaṭṭa Bhāskara and Sāyaṇa are beside the Saṃhitā for kāṇḍas 1, 2, 3 and 6. Kāṇḍas 4, 5 and 7 are being proofread.
• Timings come from forced alignment, a method that matches the known text to the audio directly. It also showed that one recording had been cut in the wrong place: we found the error by measurement, not by complaint.
• Every passage has a "Report Timing Issue" button, so listeners can correct us.

All 10,552 ṛks, each with Sāyaṇa's bhāṣya. Open the Saṃhitā as it is arranged, by maṇḍala and sūkta or by aṣṭaka and adhyāya, or gather it by who saw it, to whom it is addressed, or its metre: each of the 427 ṛṣis, 356 devatās and 82 chandas has a page of its ṛks. 1,023 sūktas carry the time each word starts, so the text keeps pace with the chant.

• Also on the site: the 1,539 samānapada mantras, on which all eight vikṛtis from jaṭā to ghana can be recited; the first 10 adhyāyas of the Aitareya Brāhmaṇa with Sāyaṇa; the Bṛhaddevatā; and indexes of every pada and of every mantra's opening words.
• The earlier timings came from an off-the-shelf tool (Aeneas). We measured them, found them wrong in places, and replaced them with our own forced alignment: 10,475 ṛks re-timed against their own audio. The check also found 163 ṛks with no usable timing, now rebuilt, and five sūktas whose audio was cut half a verse late, now re-cut.

The whole Mahābhārata, with the Harivaṃśa as its nineteenth parva, organised by parva, upaparva and adhyāya. Every verse carries Nīlakaṇṭha Caturdhara's commentary, the Bhāratabhāvadīpa. Nīlakaṇṭha compared manuscripts from many regions in the 17th century, which makes his text the first critical edition of the epic.

• 101 upaparvans, and the Harivaṃśa in its three parts: Harivaṃśa, Viṣṇu and Bhaviṣya parvas.
• 13,772 characters and places from Sorensen's Index to the Names in the Mahābhārata, each with a card.
• Facts, saṃvādas, ākhyānas and stotras, a war timeline and a map of places.
• The scanned Chitrashala volumes are linked for checking, and readers can report errors from any verse.

An atlas maps a text in three directions. Each verse is tied to the second it is chanted in an upanyāsam, a discourse in Tamil, Telugu or English where the speaker chants the Sanskrit without warning. Each verse is also tied to the people, peoples and places it names. And each of those leads to every other verse that names them, each one playable.

The Virāṭa Parva shows all three. 2,018 of its 2,262 verses link to the names in them, through 2,195 cards drawn from Sørensen's Index to the Names in the Mahābhārata. Each card carries the index's entry and the verses that name them: Kīcaka appears in 134.

The Purāṇa atlases have the first direction and not yet the other two. No index of names exists for any Purāṇa, so all 7 Purāṇa atlases have no name cards at all. Filling them is what the Purāṇa Nāma-Kośa on the funding page is for.

• 8,391 of 12,499 verses placed on the clock, from 182 recordings. How much of each text the recordings reach:

• A Śrāddha atlas over the Smṛtimuktāphalam: 208 recordings, 41.8 hours.
• Two texts have no printed edition at all. For those, the site sets two reciters against each other (Tulākāverī) or indexes the recording itself (Śrāvaṇa).

P. V. Kane's five volumes, in eight parts, as a reading edition. Every paragraph and footnote has its own address. There is a glossary, a timeline and subject indexes, and each section has a spoken synopsis.

• 171 spoken synopses across the book.
• 611 footnotes now show where the verse they quote is found. We find each one by matching Kane's own Sanskrit words, never his page numbers.
• Paragraph addresses let other sites cite a sentence, not just a file. The SMP edition below does exactly that.

Vaidyanātha Dīkṣita's digest of dharmaśāstra, in its Tamil edition, in seven kāṇḍas. The Sanskrit mūla sits beside the Tamil rendering, and the book's own contents list is the index. Every authority it quotes is searchable: Manu, Yājñavalkya, Devala and 127 more in the Śrāddhakāṇḍa alone.

• 103 links from the digest into Kane, across 50 chapters. Each one is checked on every rebuild.
• 69 places where Kane mentions the digest now open the edition.
• Transcribed from scans published by svradhakrishnasastri.in, with thanks.

Śaṅkara's commentaries on the Prasthānatraya, the ten principal Upaniṣads, the Bhagavad Gītā and the Brahmasūtra, with translations into Indian languages. Reading the bhāṣya needs deep Sanskrit; reading it beside an authoritative translation opens it to everyone else. Every verse has its own page, and the Sanskrit and its translations sit side by side.

• 13 bhāṣyas: eleven on the Upaniṣads (Kena in both its pada and vākya bhāṣyas), the Gītā and the Brahmasūtra, all with Tamil; the Īśāvāsya also in English.
• Quotations in the bhāṣya link to the passage they quote. Where the Advaita Bhāratī files lack a passage, it is taken from Sringeri's advaitasharada.sringeri.net and marked as such.

The guru-śiṣya paramparā of Advaita as a family tree, from Nārāyaṇa through Vyāsa, Gauḍapāda and Śaṅkara to the ācāryas of later centuries. Each name opens a profile with dates, an introduction and the works that ācārya wrote. A companion catalogue lists Advaita literature by prasthāna.

• 220 names in the paramparā.
• The same data prints as a wall chart for exhibitions; the full tree is below.

A chanting practice tool. The student chants a line, and the app compares it with the ācārya's recording syllable by syllable. It shows which svaras went higher or lower than they should, and lets the student hear the difference.

• Śrī Rudram and Chamakam in all four pāṭhas (saṃhitā, krama, jaṭā, ghana). That is 1,468 ghana lines for Rudram and 1,182 for Chamakam, with the ācārya's ghana class already labelled.
• We judge it with planted errors. A line is recorded with known mistakes, and the tool must catch them. It is not yet accurate enough to publish, and reaching that bar is one of the results on the funding page.